Synthesis and Optical Properties of Excited-State ... - ACS Publications

Nov 1, 2017 - namely, the major emission was from the K* (keto) form (ESIPT emission) at 515 nm with a large. Stokes shift of 11 100 cm. −1 .... tet...
26 downloads 12 Views 2MB Size
Download PDF
Article pubs.acs.org/joc

Synthesis and Optical Properties of Excited-State Intramolecular Proton Transfer Active π‑Conjugated Benzimidazole Compounds: Influence of Structural Rigidification by Ring Fusion Koji Takagi,*,† Kaede Ito,† Yoshihiro Yamada,† Takuya Nakashima,‡ Ryoichi Fukuda,§,∥ Masahiro Ehara,§,∥ and Hyuma Masu⊥ †

Graduate School of Engineering, Nagoya Institute of Technology, Gokiso-cho, Showa-ku, Nagoya 466-8555, Japan Graduate School of Materials Science, Nara Institute of Science and Technology, 8916-5 Takayama, Ikoma, Nara 630-0192, Japan § Research Center for Computational Science and Institute for Molecular Science, 38 Nishigo-naka, Myodaiji, Okazaki 444-8585, Japan ∥ Elements Strategy Initiative for Catalysts and Batteries, Kyoto University, 1-30 Goryo-Ohara, Nishikyo-ku, Kyoto 615-8245, Japan ⊥ Center for Analytical Instrumentation, Chiba University, 1-33 Yayoi, Inage, Chiba, Chiba 263-8522, Japan ‡

S Supporting Information *

ABSTRACT: Two excited-state intramolecular proton transfer (ESIPT) active benzimidazole derivatives (1 and 2) were synthesized by acid-catalyzed intramolecular cyclization. The steadystate fluorescence spectrum in THF revealed that ring-fused derivative 1 exhibits a dual emission, namely, the major emission was from the K* (keto) form (ESIPT emission) at 515 nm with a large Stokes shift of 11 100 cm−1 and the minor emission was from the E* (enol) form at below 400 nm. In contrast, the normal emission from the E* form was dominant and the fluorescence quantum yield was very low (Φ ∼ 0.002) for nonfused derivative 2. The time-resolved fluorescence spectroscopy of 1 suggested that ESIPT effectively occurs due to the restricted conformational transition to the S1−TICT state, and the averaged radiative and nonradiative decay rate constants were estimated as ⟨kf⟩ = 0.15 ns−1 and ⟨knr⟩ = 0.60 ns−1, respectively. The fluorescence emission of 1 was influenced by the measurement conditions, such as solvent polarity and basicity, as well as the presence of Lewis base. The ESIPT process and solvatochromic behavior were nicely reproduced by the DFT/TDDFT calculation using the PCM model. In the single-crystal fluorescent spectra, the ESIPT emissions were exclusively observed for both fused and nonfused compounds as a result of hydrogen-bonding interactions.

■

INTRODUCTION Because of the diverse molecular design for the on-demand application, a large number of fluorescent π-conjugated compounds have been developed. However, most of them have serious problems, even if they are highly fluorescent in dilute solutions. In the solid state, the molecular aggregation and intermolecular contact bring about the nonradiative decay from the excited state, resulting in the fatal decrease of the fluorescence intensity or the complete fluorescence quenching.1,2 In the field of optoelectronic device applications, on the other hand, the efficient fluorescence emission in the solid state is extremely desirable. One promising strategy is based on the excited-state intramolecular proton transfer (ESIPT) fluorophores.3,4 ESIPT-active molecules generally adopt the stable enol (E) form in the ground state and the transformation to the keto (K*) form promptly occurs upon photoexcitation, although this is not always the case. The formation of intramolecular hydrogen bonding often facilitates this so-called four-level cyclic proton-transfer process (E−E*−K*−K−E).5−9 The absorption by the E form and the emission from the K* form induce a large Stokes shift that is the energy difference © 2017 American Chemical Society

between the absorption and emission maxima, which overcomes the fluorescence self-quenching even in the solid state and achieves the bright light emission. Novel oxazole-, thiazole-, and imidazole-based fluorescent compounds harnessing the ESIPT mechanism have been synthesized. In particular, imidazole derivatives are promising candidates among them. Park et al. have reported ESIPT-active π-conjugated compounds based on the 2-(2′-hydroxyphenyl)imidazole unit.10−14 Mutai et al. have prepared structurally relevant 2-(2′hydroxyphenyl)imidazo[1,2-a]pyridines to reveal the color tunability due to the substituent effect15,16 and polymorphdependent multicolor luminescence.17,18 Gryko and colleagues have synthesized a set of π-extended ESIPT-active 2-(2′hydroxyphenyl)imidazole derivatives19 and 2-(2′-arylsulfonamidophenyl)imidazole derivatives20 to investigate the structure−photophysical property relationship. On the other hand, the formation of intramolecular hydrogen bonding is not a sufficient condition for the ESIPT emission. Received: August 4, 2017 Published: November 1, 2017 12173

DOI: 10.1021/acs.joc.7b01967 J. Org. Chem. 2017, 82, 12173−12180

Article

The Journal of Organic Chemistry

time-resolved fluorescence spectra of 1 and 2 as well as a previously studied reference compound 1′ without the hydroxy group were measured in solutions. The obtained results were discussed on the basis of the quantum chemical calculations with the DFT and time-dependent (TD) DFT methods using the PBE0 functional. The single-crystal fluorescence spectra were also measured to reveal the emission characteristic in the solid state. There is room for discussion about the relationship between the packing structure and fluorescence characteristics of benzimidazole-based ESIPT-active molecules, and the results obtained here give useful information for the solid-state lightemitting materials.

Park and co-workers21 and Tsai et al.22 have performed the theoretical calculation and investigated the potential energy surface to discuss the first singlet excited state behavior of ESIPT-active molecules. As a result, stable excited-state twisted intramolecular charge transfer (S1−TICT) states responsible for the nonradiative relaxation were found to disturb the fluorescence emission from the K* form.22 Because the intramolecular charge transfer (ICT) state can be formed by the internal rotation around the C (imidazole)−C (phenol) single bond, the confinement of proton donor and acceptor segments into a planar rigid conformation would be a straightforward method to inhibit the transition to the undesirable S1−TICT state. To this end, several ESIPT-active polyacenes, including 10-hydroxybenzo[h]quinoline,23,24 10aminobenzo[h]quinoline,25 and 12-hydroxy-1-azaperylene,26 were prepared and their photophysical properties were carefully examined. During the course of our investigation on benzimidazole-based fluorescent molecules27 incorporating the ESIPT unit, three papers describing the efficient ESIPT emission were published. Hung and Chou et al. have synthesized locked ortho- and para-core chromophores of green fluorescent protein to figure out that the structural constraint is effective for the high emission yield from the keto tautomer.28 Parada et al. have investigated the impact of hydrogen-bond geometry on the dynamics of ESIPT and revealed the importance of the proton donor−acceptor distance as well as the dihedral angle between the proton donor− acceptor subunits.29 In close relation to our study, Skonieczny et al. have demonstrated the use of molecular design to reach intense luminescence for compounds capable of ESIPT using a family of 2-(2′-hydroxyphenyl)benzimidazole derivatives.30 Herein, two benzimidazole derivatives (Figure 1, 1 and 2) carrying a hydroxy group amenable to form an intramolecular

■

RESULTS AND DISCUSSION Synthesis. Fused π-conjugated compound 1 was synthesized from 2-(4′-methoxybenzyloxy)-3,4-dimethoxybenzaldehyde31 in three steps [Scheme S1, Supporting Information (SI)], namely, the imidazole ring formation (A), N-alkylation with bromoacetaldehyde diethyl acetal (B), and TfOHcatalyzed intramolecular cyclization. In the final step, the deprotection of the phenolic hydroxy group simultaneously occurred to obtain 1 in a moderate yield. Concentrated HCl solution was insufficient to bring about the desired reaction. Compound 2 without the bridging CC double bond between the imidazole and phenol units was synthesized by the deprotection of the phenolic hydroxy group of the intermediate product A using CeCl3 and NaI (Scheme S2, SI). The structure and purity of compounds 1 and 2 were validated by the highresolution NMR spectra and elemental analyses. Optical Properties in Solution. The ultraviolet (UV) absorption and fluorescence emission spectra of 1, 1′, and 2 in tetrahydrofuran (THF) are indicated in Figure 2 and the data

Figure 1. Structure of benzimidazole derivatives with the hydroxy group (1 and 2) and without the hydroxy group (1′). The structure of compound 1″ with the C−C single bond (not described in this paper) is also shown.

Figure 2. (A) UV spectra and (B) fluorescence spectra of 1 (red line), 1′ (green line), and 2 (blue line) in THF (10−5 M).

hydrogen bond were synthesized. In our previous paper,27 we have synthesized a benzimidazole analog (1′) without the hydroxy group. The aim of this work is to address the influence of the hydroxy group as well as the structural rigidification on photophysical (ESIPT) properties. Meanwhile, 1″ with the C− C single bond as the strap may be another candidate. The preliminary density functional theory (DFT) calculation at the B3LYP/6-31G level of theory suggested the increased dihedral angle (7°) of 1″ between the proton donor−acceptor subunits as compared with that of 1 (1°) (detailed data are not shown here). Because a large dihedral angle is responsible for the nonradiative deactivation channel after the ESIPT process and low fluorescence quantum yield,29 we have chosen 1 instead of 1″ to realize the high K* emission yield. The steady-state and

are summarized in Table S1 (SI). All UV spectra showed vibronic fine structures to indicate the rigid conformation in the ground state. The UV spectrum of 1 (Figure 2A, red line, λab = 328 nm) exhibited a hypsochromic shift by 10 nm from that of 1′ (Figure 2A, green line, λab = 339 nm), which was investigated on the basis of the theoretical calculation. The DFT calculations revealed that 1 adopts a conformation that deviated from the planar structure due to the out-of-plane orientation of methoxy groups (Figure S11, SI), and its first excitation energy of 3.86 eV (321 nm, f = 0.12) calculated by TDDFT is higher compared with that of 1′ having the planar ground-state structure (3.75 eV, f = 0.18) (Table S6, SI). They have a π−π* excitation characteristic with relatively high intensity of oscillator strength ( f). This theoretical analysis well 12174

DOI: 10.1021/acs.joc.7b01967 J. Org. Chem. 2017, 82, 12173−12180

Article

The Journal of Organic Chemistry

transfer. The relative peak intensity of the E* emission/K* emission was 0.16/1. The time-resolved fluorescence decay measurement of 1 was carried out in THF with the excitation wavelength of 365 nm and the monitoring wavelengths of 500 and 400 nm (Figure S5, SI). The fluorescence decay profile monitored at 500 nm showed a dual-exponential decay with time constants of 1.30 ns (97%) and 6.02 ns (3%), giving the averaged lifetime of 1.34 ns. Together with the fluorescence quantum yield (Φ = 0.20), the averaged rate constants for radiative and nonradiative decays are estimated as ⟨kf⟩ = 0.15 ns−1 and ⟨knr⟩ = 0.60 ns−1, respectively (Table S2, SI). These values are also roughly comparable to those of the 2-hydroxyphenyl-functionalized phenanthro[9,10-d]imidazole derivative.32 Meanwhile, the decay profile observed at 400 nm corresponding to the emission from the LE state is coincident with the instrument response function (IRF) profile (Figure S5, SI), suggesting the ESIPT process with a time constant far less than 25 ps. Thus, the ESIPT process is relatively fast, but room remains for argument by comparing the values reported before.28,29,33,34 Our instrumental setup makes an accurate kinetic analysis at the early stage (on the order of picosecond or subpicosecond) after the photoexcitation difficult and results in a rough estimation of the time constant. The kinetic analysis in a cyclohexane solution likely gives invaluable data, because the overall reaction dynamics can be described by a mechanism incorporating both solvent polarization and the proton transfer reaction coordinate.35 However, the compound is poorly soluble in nonpolar hydrocarbon solvents, and also the timeresolved fluorescence measurement using our apparatus has an IRF of around several tens of picoseconds. Accordingly, the fact we can claim at the present stage is that the ESIPT process is faster than 25 ps. The excitation spectrum monitored at 515 nm was almost identical with the absorption spectrum (Figure S6, SI). Those features agree well with the DFT/TDDFT results. The K form of 1 in the first excited state (K*) is calculated to be far more stable than the E* form by 0.65 eV, indicating the possibility of ESIPT; the details will be discussed in the following subsection. The calculated fluorescence energies of 1 in THF are 3.21 eV (386 nm) and 2.10 eV (591 nm) from the E* and K* forms, respectively. The fluorescence energy from the E* form is close to that of 1′ calculated at 3.42 eV (362 nm) (Table S6 and Figure S12, SI). The UV spectrum of 2 was observed at the shortest wavelength (Figure 2A, blue line, λab = 318 nm) among three compounds due to the absence of the conjugated CC double bond. The fluorescence spectrum consisted of two bands for which the intensity of E* emission/K* emission was 1/0.64. The shorter wavelength fluorescence, having a peak maximum at 348 nm (Figure 2B, blue line), can be assigned to the emission from the E* form because this emission band is a mirror image of the absorption band and the Stokes shift was calculated as 2700 cm−1. The longer wavelength fluorescence at 493 nm, having a Stokes shift of 11 200 cm−1, most likely originated from the K* form. The total fluorescence quantum yield was very low (Φ ∼ 0.002), probably due to the nonradiative decay through the S1−TICT state.21,22 The excitation spectrum of 2 monitored at 493 nm exhibited a different peak pattern from the absorption spectrum, but an accurate discussion was difficult due to the quite low fluorescence quantum yield. The time-resolved fluorescence decay measurement of 2 was carried out in THF with the excitation wavelength of 350 nm and the monitoring

reproduced the experimental result. The absorption and emission characteristics of 1′ were almost identical with those of benzo[4,5]imidazo[1,2-f ]phenanthridine reported by Skonieczny and Gryko.32 The steady-state fluorescence spectrum of 1′ showed a vibronic fine structure, and the peak maximum was observed at 377 nm (Figure 2B, green line). The relative fluorescence quantum yield using quinine sulfate as a standard was 0.32. The absorption and emission spectra were a mirror image of each other, and the Stokes shift was calculated as 3000 cm−1. These facts suggest that the fluorescence is obtained from the localized excited (LE) state. The time-resolved fluorescence decay measurement of 1′ was carried out in THF with the excitation wavelength of 365 nm and the monitoring wavelength of 400 nm. As a result, the single exponential decay from the singlet excited state was obtained (Figure S5, SI). From the fluorescence quantum yield (Φ = 0.32) and the fluorescence lifetime (τ = 3.62 ns), the radiative and nonradiative decay rate constants of 1′ were calculated as kf = 0.09 and knr = 0.19 ns−1, respectively (Table S2, SI). On the other hand, 1 exhibited a dual emission having a major peak at 515 nm and a minor peak below 400 nm (Figure 2B, red line). The total fluorescence quantum yield (Φ = 0.20) was smaller than that of 1′ and comparable to that of 2-hydroxyphenylfunctionalized phenanthro[9,10-d]imidazole derivatives.10,19 The longer wavelength fluorescence with a large Stokes shift of 11 100 cm−1 can be ascribed to the emission from the K* form as a result of the ESIPT process (Scheme 1). An efficient Scheme 1

ESIPT emission observed for 1 would be ascribed to the restricted rotation around the C (imidazole)−C (phenol) single bond (vide infra) as well as the fact that the planar conformation is most stable in 5-membered-ring ESIPT molecules.21 The shorter wavelength fluorescence is supposed to originate from the E* form because the weak emission bands at 362, 382, and 400 nm roughly accord with the LE state emission from 1′ at 358, 377, and 398 nm. Although an intramolecular hydrogen bond between the CN and H−O groups is actually formed, as judged from the single crystal Xray structure (vide infra, Figure 6), two electron-donating methoxy substituents on the phenol ring likely decrease the acidity of the hydroxy proton and cause the incomplete proton 12175

DOI: 10.1021/acs.joc.7b01967 J. Org. Chem. 2017, 82, 12173−12180

Article

The Journal of Organic Chemistry

properties of 1 measured in DMF solution. Likewise, the steady-state fluorescence emission spectra of 2 in solutions were collected. Irrespective of the solvent character, the emissions from the E* form were always predominantly observed (Figure S9, SI). Particularly, in MeOH solution, the ESIPT fluorescence was very weak, likely due to the breaking of the intramolecular hydrogen bond. The fluorescence quantum yield was still low in MeOH (Φ = 0.02) but increased from that in THF solution (Φ < 0.01). The above assumptions for the solvatochromic behavior were confirmed by the DFT/TDDFT calculations using the polarizable continuum model (PCM). In methanol, the calculated excitation energies are almost same as those in THF, while the calculated fluorescence energy in methanol exhibits a hypsochromic shift by 0.11 eV (30 nm) in comparison with that in THF (Tables S5−S8, SI). This fluorescence shift in polar solvent agrees with the experimental findings, although our calculations did not consider hydrogen bonds explicitly. The lowest excitation energy of the deprotonated anion (A) form of 1 is calculated as 2.73 eV (454 nm, f = 0.29) in THF; thus, the absorption of 1 would exhibit a bathochromic shift under basic conditions. This result agrees well with the experimental findings. The fluorescence from A* is calculated as 2.33 eV (532 nm). These computational results strongly support that the new fluorescence band observed in DMF or in THF with excess TEA is ascribed to the emission from the A* form (Tables S9 and S10, SI). Quantum Chemical Calculation. The absorption and emission behavior through the intramolecular hydrogenbonding-assisted ESIPT process of 1 was investigated in detail with DFT/TDDFT calculations. Figure 4 shows the simulated

wavelength of 500 nm (Figure S6, SI). Although the low fluorescence quantum yield made an accurate analysis difficult, the radiative and nonradiative decay rate constants of 2 were roughly estimated as kf = 0.002 and knr = 1.18 ns−1, respectively (Table S2, SI). These results are in sharp contrast to those of 1 and definitely suggest that the structural rigidification due to the ring fusion significantly influences the excited state geometry and dynamics. Subsequently, the steady-state fluorescence emission spectra of 1 in solutions were collected under various conditions. In contrast to the similar absorption spectra (Figure S7, SI), in which the π−π* transition peaks are always located at around 330 nm, the corresponding fluorescence emission revealed obvious differences depending on the solvent character. The emission from the E* form markedly descended and the ESIPT emission was dominant in dichloromethane (DCM) solution (Figure 3A, orange line). The fluorescence quantum yield was

Figure 3. (A) Fluorescence spectra of 1 in THF (blue line), DCM (orange line), DMF (green line), and MeOH (black line) (10−5 M). (B) Fluorescence spectra of 1 in THF without (black line) and with (blue line) an excess amount of TEA (100 equiv relative to 1).

decreased from THF solution (Φ = 0.20) to DCM solution (Φ = 0.06). In methanol (MeOH) solution, the emission from the E* form was increased and the emission from the K* form exhibited a hypsochromic shift to 486 nm (black line). These results suggest that the competitive hydrogen bond formation with MeOH is the origin of the increased E* emission and that the polar solvent destabilizes the singlet excited state and/or stabilizes the ground state of the keto forms.30 On the other hand, a new emission band at 416 nm was observed in addition to the emission from the K* form in N,N-dimethylformamide (DMF) solution (green line), and the total fluorescence quantum yield was calculated as Φ = 0.27. As previously described,36,37 it can be conceived that 1 is deprotonated by DMF, giving rise to its anionic form. In the absorption spectrum of 1 in basic DMF solution, actually a small but unambiguous absorption band ranging from 360 to 400 nm ascribable to the anionic form was observed (Figure S7, SI), which is in good agreement with the result reported by Mosquera et al.38 In order to gain deeper insight into the influence of DMF, the UV absorption and fluorescence emission spectra of 1 were measured by adding an excess amount of basic triethylamine (TEA, 100 equiv relative to 1) to a THF solution. A new absorption band appeared at 380 nm in the UV spectrum (Figure S8, SI), while a large emission band having a peak maximum at 432 nm was observed in the fluorescence spectrum (Figure 3B). The fluorescence quantum yield was increased (Φ = 0.41) from the value obtained in the absence of TEA. These results nicely reproduced the optical

Figure 4. Simulated absorption and fluorescence spectra obtained by the DFT/TDDFT calculations for 1 in the E (red), K (blue), and A (green) forms in THF. The spectrum was simulated by a Gaussian distribution for which the full width at half-maximum (fwhm) is set to 0.2 eV.

absorption and fluorescence spectra of 1 for E, K, and A forms. The computations did not provide absorption bands lower than 350 nm for the E form. We found that the first absorption band obtained around 300 nm is composed of two excited states. The K form exhibits lower-energy absorption bands than the E form by 0.88 eV (95 nm). The first excitation band of the A form is calculated in the lower-energy region. The UV absorption of the E form is thus clearly distinguished from those of the K and A forms. For the fluorescence spectra, the emission band of the A form is calculated to be between those 12176

DOI: 10.1021/acs.joc.7b01967 J. Org. Chem. 2017, 82, 12173−12180

Article

The Journal of Organic Chemistry

Optical Properties in Single Crystal. For determining and comparing the ESIPT performance of 1 and 2, the fluorescence characteristic was evaluated in single crystals. Colorless single crystals of 1 and 2 suitable for X-ray crystallographic analysis were grown from hexane/CHCl3 and CH2Cl2/MeOH, respectively. The thermal ellipsoidal models as well as the packing structures are provided in Figure 6. The

of the E and K forms. The calculated Stokes shift of the A form is 0.40 eV (3230 cm−1), which is similar to that of 1′. Figure 5 shows the calculated energy diagram of 1 in THF relative to the E form in the ground (S0) state. The energies of

Figure 5. Calculated energy diagram (in eV) related to the ESIPT fluorescence emission of 1. The transition energies are given in parentheses.

Figure 6. Thermal ellipsoidal models (left) and packing structures (right) of (A) 1 and (B) 2. The ellipsoids are drawn at 50% probability level, while isotropic hydrogen atoms are represented by spheres of arbitrary size. Hydrogen bonds are indicated by sky-blue lines.

the first excited state (S1) were estimated by the transition (absorption and fluorescence) energies. The relative energy of the A form in the ground state was estimated using the calculated protonation energy of TEA assuming 1(E) + Et3N → 1(A) + Et3NH+ + ΔE, where ΔE denotes the deprotonation energy of 1 and also the relative energy of the A form (this value seems to be not so accurate, but it does not affect the following discussion). In the S0 state, the K form is less stable by 0.41 eV than the E form; therefore, we may consider that almost all of 1 in solution takes the E form. In the S1 state, the K form is remarkably stabilized, and ESIPT is supposed to occur by an exothermic reaction. The relaxation energy is calculated to be 1.0 eV between the vertical excitation of the E form and the relaxed geometry in the K* form. Solvent molecules likely affect the energy barrier associated with the geometry transition in the excited state. On the other hand, the picosecond kinetic studies suggest that the excited-state proton transfer is essentially barrier-free in noninteracting solvents.9 Our theoretical calculation may include a certain level of error and does not fully account for the detailed kinetics in the excited state. Although the potential energy surface calculation would be helpful for the in-depth discussion, this theoretical analysis is beyond the scope of this study and will be investigated in the future. Meanwhile, our PCM calculations in methanol show that the K form in the S0 state gains relatively large solvation energy. This stabilization of the S0 state in the K form is the origin of the hypsochromic shift for the fluorescence in methanol. The effect of the hydrogen bond could not be elucidated by the present model. In the A form, the S0 state is destabilized by the negative charge. In short, the excitation energy is lowered in the anion because the HOMO (highest occupied molecular orbital) level rises. On the other hand, the geometrical relaxation in the S1 state is small (0.28 eV), and therefore, its fluorescence energy becomes relatively high. Consequently, the fluorescence band of the A form is observed between those of the E and K forms.

detailed crystal data are summarized and the full-size images are provided in Figures S13 and S14 (SI). It was found that the molecules of 1 (monoclinic, space group P21/n, Z = 4) are stacked along the columnar axis in an antiparallel fashion with the π-stacking distance of d = 3.43 Å. The torsion angle between hydroxyphenyl and benzimidazole rings is as small as θ = 2.2°, and the distance between O−H···N (d = 2.71 Å) is less than the common O−H···N length (d = 2.80 Å). These results definitely indicate that 1 adopts the coplanar conformation assisted by the intramolecular hydrogen bonding. As for 2 (monoclinic, space group P21/c, Z = 4), on the other hand, the molecules have a more twisted conformation, as judged by the torsion angle between the hydroxyphenyl and benzimidazole rings of θ = 9.3°. The short distance between O−H···N (d = 2.58 Å) supports the formation of an intramolecular hydrogen bond. It should be emphasized that the molecules of 2 do not π-stack but give a linear hydrogen-bonding network structure where each molecule is arranged in a perpendicular fashion with two kinds of intermolecular N−H···O hydrogen bonding (d = 2.21 and 2.61 Å) (See Figure S15, SI). The single-crystal fluorescence spectra were measured by microscopic spectroscopy, and the results are listed in Table S5 (SI). Figure 7A demonstrates that, upon the excitation with UV light (330 nm < λex < 385 nm), the single crystal of 1 exclusively emitted ESIPT fluorescence, having a peak maximum at 484 nm (red line). The absolute fluorescence quantum yield was estimated as Φ = 0.16. Because the molecules in the single crystal are forced to adopt the planar conformation (vide supra), which accelerates the phenolic proton to transfer to the proton-accepting imidazole nitrogen, no emission band originating from the E* form was observed at all. Neither the excimer emission at the longer wavelength region nor the detrimental drop of the fluorescence quantum yield was observed, implying that the molecules in the excited state have negligible electronic interactions with the neighbor12177

DOI: 10.1021/acs.joc.7b01967 J. Org. Chem. 2017, 82, 12173−12180

Article

The Journal of Organic Chemistry 13

C-nuclear magnetic resonance (NMR) spectroscopies were obtained on a Bruker Avance III HD NanoBay 400 FT-NMR spectrometer using tetramethylsilane (1H NMR) and solvent residual peaks (13C NMR) as the internal standards. Melting points (Mp) were determined on a Yanaco micro melting point apparatus (MP-500D). High-resolution electrospray ionization mass spectra (HR ESI-MS) were obtained on a Waters Synapt G2 HDMS. Elemental analyses (EA) were performed on a Elementar vario EL cube in the CHN mode. Ultraviolet−visible (UV−vis) and fluorescence spectra in solutions were recorded on a Shimadzu UV-1650 spectrophotometer and a Shimadzu RF-5300 spectrofluorometer, respectively, using a 10 mm quartz cell. Fluorescence quantum yields (QYs) were determined relative to quinine sulfate in 0.1 M H2SO4 having a QY of 0.55. Photoluminescence lifetime measurements in solutions were performed using a picosecond fluorescence measurement system (Hamamatsu C4780) with a streak scope (Hamamatsu C4334). The excitation source was generated by a Nd:YVO4 laser (Coherent, Verdi) pumped Ti:sapphire laser system (Coherent, Mira-900) equipped with a cavity dumper (Coherent, PulseSwitch). This delivers 100 fs pulse trains at 730 nm (or 700 nm) and runs at a repetition rate of 5.43 kHz for the time-resolved photoluminescence measurements. After the frequency was doubled (λex = 365 or 350 nm) with a LiB3O5 (LBO) crystal, the incident pulses were focused on the sample. Fluorescence spectra in crystalline powders were recorded on a Hamamatsu Quantaurus-QY Plus C13534 instrument and absolute QYs were obtained with an integrating sphere C11347. Single crystals were irradiated by a mercury lamp in an epi-fluorescence attachment of a polarizing microscope (Olympus BX-51) through an excitation filter (330 nm < λex < 385 nm), and the fluorescence through a long pass filter (λ > 400 nm) was detected with a Hamamatsu PMA-11 photodetector. Absolute QYs were obtained using a Hamamatsu C9920-02 with an integrating sphere. Theoretical calculations were performed using the Gaussian 09 (Revision E.01) package of programs.39 The ground-state structure was optimized with DFT calculation at the PBE040/cc-pVDZ41−43 level of theory. The TDDFT calculation was then performed at the PBE0/aug-cc-pVDZ (for C, N, O) or cc-pVDZ (for H) level to estimate the vertical excitation energies and oscillator strengths. The solvent effect was considered by the IEF PCM44−46 (integral equation formula polarizable continuum model). For electronic excitation from the ground state (absorption spectrum), the linear response formalism47 was used, while for emission from the first excited state (fluorescence spectrum), the statespecific and nonequilibrium solvation formalism48 was used. Crystallographic data were collected on a CCD diffractometer with Cu Kα (λ = 1.54178 Å) radiation. Data collections were carried out at low temperature (173 K) using liquid nitrogen. All of the crystal structures were solved by direct methods with SHELXS-97 and refined with fullmatrix least-squares SHELXL-2013.49 All non-hydrogen atoms were refined anisotropically, and hydrogen atoms were included at their calculated positions. 2-(2′-(4″-Methoxybenzyl)-3′,4′-dimethoxyphenyl)benzimidazole (A). A mixture of 2-(4′-methoxybenzyloxy)-3,4-dimethoxybenzaldehyde (4.47 g, 15 mmol),39 o-phenylenediamine (1.33 g, 12 mmol), and Cu(OH)2 (120 mg, 1.2 mmol) in MeOH (36 mL) was vigorously stirred under an oxygen atmosphere at room temperature for 6.5 h. After MeOH (237 mL) was added, Cu(OH)2 was removed by the filtration. Solvents were evaporated, and the crude product was purified by SiO2 column chromatography (acetone/hexane = 1/1, Rf = 0.7) to obtain 2.94 g of a pale yellow solid. 1H NMR (DMSO-d6): δ ppm 12.02 (s, 1H), 7.79 (d, J = 8.8 Hz, 1H), 7.62−7.57 (m, 1H), 7.54−7.48 (m, 1H), 7.24 (d, J = 8.6 Hz, 2H), 7.20−7.14 (m, 2H), 6.98 (d, J = 8.8 Hz, 1H), 6.77 (d, J = 8.6 Hz, 2H), 5.05 (s, 2H), 3.88 (s, 3H), 3.85 (s, 3H), 3.67 (s, 3H). HRMS (ESI/TOF-Q) m/z: [M + H]+ calcd for C23H22N2O4 391.1658, found 391.1658. Because a complete purification was impossible, this material was used as a mixture including byproducts for the following reaction, and 13C NMR data are not provided. 1-(1′,1′-Diethoxyethyl)-2-(2″-(4‴-methoxybenzyl)-3″,4″-dimethoxyphenyl)benzimidazole (B). A mixture of A (2.94 g, 7.5 mmol), K2CO3 (2.08 g, 15 mmol), and bromoacetaldehyde diethyl

Figure 7. Fluorescence spectra of (A) 1 and (B) 2 in THF solution (black line) and a single crystal (red line).

ing enol molecules in the ground state in spite of the closely πstacked structure.15 Figure 7B clearly shows the big change in the fluorescence spectra of 2 by changing from THF solution to a single crystal. Namely, only the ESIPT emission was observed at 440 nm. It should be noted that the emission profile was not measured appropriately below 400 nm because of the emission long pass filter in the microscope, while the absence of E* form emission was definitely confirmed by the steady-state fluorescence measurement for the crystalline powder sample (Figure S10, SI). It is conceivable that not only the intramolecular but also the intermolecular hydrogen-bonding interaction makes the proton transfer process easier, although further experimental work is needed to get a conclusion. It is quite important that the fluorescent quantum yield was dramatically increased from THF solution (Φ = 0.002) to a single crystal (Φ = 0.42). Because the twisting motion and the related conformational transition to the S1−TICT state are difficult in the crystal state, the nonradiative decay from the singlet excited state is likely prohibited, even if 2 has a flexible conformation in the solution state.

■

CONCLUSIONS We have synthesized ring-fused and nonfused ESIPT-active 2(2′-hydroxyphenyl)benzimidazole derivatives through acidcatalyzed intramolecular cyclization. In the solution-state fluorescent spectrum, the ring-fused compound mainly exhibited an ESIPT emission from the K* form having a large Stokes shift, while the fluorescence emission from the E* form was dominant for the nonfused variant. The fluorescence quantum yield was considerably increased in the fused πconjugated imidazole derivative most likely by blocking the C− C bond rotation around the hydroxyphenyl and imidazole units. The emission maximum wavelength, as well as the fluorescence quantum yield of the ring-fused compound, was influenced by the solvent character. The ESIPT process and solvatochromic behavior can be well-explained by the highly accurate quantum chemical calculations. In the single-crystal fluorescent spectra, the ESIPT emissions were exclusively observed by adopting the completely planar structure (fused compound) and the formation of an intermolecular hydrogen bond (nonfused compound).

■

EXPERIMENTAL SECTION

Materials and Instruments. All materials were obtained from commercial suppliers and used without purification unless otherwise noted. Anhydrous tetrahydrofuran (THF) was purchased from Kanto Chemical Co. Other solvents were dried and distilled, following standard methods, and stored under a nitrogen atmosphere. 1H- and 12178

DOI: 10.1021/acs.joc.7b01967 J. Org. Chem. 2017, 82, 12173−12180

The Journal of Organic Chemistry acetal (1.72 mL, 11 mmol) in dimethyl sulfoxide (DMSO, 66 mL) was heated at 120 °C overnight. After dichloromethane (DCM) and water were added, the aqueous phase was extracted with DCM. The combined organic phase was dried over MgSO4, solvents were evaporated, and the crude product was purified by SiO2 column chromatography (acetone/hexane = 1/2, Rf = 0.5) to obtain 1.83 g of a brown oil. 1H NMR (DMSO-d6): δ ppm 7.69−7.61 (2H), 7.26 (H), 7.02 (d, J = 8.8 Hz, 1H), 6.88 (d, J = 8.6 Hz, 2H), 6.62 (d, J = 8.6 Hz, 2H), 4.73 (s, 2H), 4.52 (t, J = 5.3 Hz, 1H), 4.05 (d, J = 5.4 Hz, 2 H), 3.89 (s, 3H), 3.83 (s, 3H), 3.64 (s, 3H), 3.08 (brs, 2H), 0.83 (t, J = 7.0 Hz, 6H). HRMS (ESI/TOF-Q) m/z: [M + H]+ calcd for C29H34N2O6 507.2495, found 507.2495. Because a complete purification was impossible, this material was used as a mixture including byproducts for the following reaction, and 13C NMR data are not provided. Synthesis of 1. A DCM solution (72 mL) of B (1.83 g, 3.6 mmol) and TfOH (7 mL, 79 mmol) was stirred at room temperature for 1.5 h. After DCM and water were added, the aqueous phase was extracted with DCM. The combined organic phase was dried over MgSO4, solvents were evaporated, and the crude product was purified by SiO2 column chromatography (acetone/hexane = 1/2, Rf = 0.4) to obtain 532 mg of a colorless solid (50% yield). Mp: 209.7−210.7 °C. 1H NMR (DMSO-d6): δ ppm 11.96 (brs, 1H), 8.73 (d, J = 7.3 Hz, 1H), 8.27 (d, J = 8.1 Hz, 1H), 7.89 (d, J = 8.1 Hz, 1H), 7.55−7.49 (m, 1H), 7.48−7.41 (m, 1H), 7.27 (d, J = 7.3 Hz, 1H), 7.13 (s, 1H), 3.95 (s, 3H), 3.86 (s, 3H). 13C NMR (CDCl3): δ ppm 156.0, 150.3, 147.4, 142.2, 135.9, 128.6, 128.0, 124.7, 121.6, 120.3, 118.9, 111.5, 109.6, 105.8, 99.4, 60.9, 56.1. Anal. Calcd for C17H14N2O3: C, 69.38%; H, 4.79%; N, 9.52%. Found: C, 69.07%; H, 4.67%; N, 9.35%. Synthesis of 2. A CH3CN solution (47 mL) of A (1.00 g, 2.6 mmol), CeCl3·7H2O (1.42 g, 3.8 mmol), and NaI (384 mg, 2.6 mmol) was heated to reflux for 42 h. After 1 M HCl (12 mL) and DCM were added, the aqueous phase was extracted with DCM. The combined organic phase was dried over MgSO4, solvents were evaporated, and the crude product was purified by SiO2 column chromatography (acetone/hexane = 1/2, Rf = 0.4) followed by recrystallization from DCM/MeOH to obtain 147 mg of a colorless solid (21% yield). Mp: 127.4−128.7 °C. 1H NMR (DMSO-d6): δ ppm 13.41 (brs, 1H), 13.10 (brs, 1H), 7.77 (d, J = 9.0 Hz, 1H), 7.67 (d, J = 6.8 Hz, 1H), 7.56 (d, J = 7.1 Hz, 1H), 7.26 (2H), 6.77 (d, J = 9.0 Hz, 1H), 3.86 (s, 3H), 3.76 (s, 3H). The 13C NMR spectrum was not available due to its poor solubility. Anal. Calcd for C15H14N2O3·0.3H2O: C, 65.35%; H, 5.34%; N, 10.16%. Found: C, 65.34%; H, 5.35%; N, 9.97%.

■

■

ACKNOWLEDGMENTS

■

REFERENCES

A part of this work was technically supported by the Nanotechnology Platform Program (Molecule and Material Synthesis) of the Ministry of Education Culture, Sports, Science and Technology (MEXT), Japan. A part of this work was financially supported by and performed under the Cooperative Research Program of “Network Joint Research Center for Materials and Devices: Dynamic Alliance for Open Innovation Bridging Human, Environment and Materials”. R.F. acknowledges the financial support from a Grant-in-Aid for Scientific Research from the Japan Society for the Promotion of Science (JSPS) (26410026). The computations were partially performed at the Research Center for Computational Science, Okazaki, Japan. We also thank Mr. Y. Okajima at NAIST for his assistance with fluorescence lifetime measurements.

(1) Birks, J. B. Photophysics of Aromatic Molecules; Wiley: New York, 1970. (2) Ding, D.; Li, K.; Liu, B.; Tang, B. Z. Acc. Chem. Res. 2013, 46, 2441−2453. (3) Padalkar, V. S.; Seki, S. Chem. Soc. Rev. 2016, 45, 169−202. (4) Zhao, J.; Ji, S.; Chen, Y.; Guo, H.; Yang, P. Phys. Chem. Chem. Phys. 2012, 14, 8803−8817. (5) Sengupta, P. K.; Kasha, M. Chem. Phys. Lett. 1979, 68, 382−385. (6) McMorrow, D.; Kasha, M. J. Am. Chem. Soc. 1983, 105, 5133− 5134. (7) McMorrow, D.; Kasha, M. J. Phys. Chem. 1984, 88, 2235−2243. (8) Dick, B.; Ernsting, N. P. J. Phys. Chem. 1987, 91, 4261−4265. (9) Barbara, P. F.; Walsh, P. K.; Brus, L. E. J. Phys. Chem. 1989, 93, 29−34. (10) Park, S.; Kwon, J. E.; Park, S. Y. Phys. Chem. Chem. Phys. 2012, 14, 8878−8884. (11) Park, S.; Kwon, O.-H.; Lee, Y.-S.; Jang, D.-J.; Park, S. Y. J. Phys. Chem. A 2007, 111, 9649−9653. (12) Park, S.; Seo, J.; Kim, S. H.; Park, S. Y. Adv. Funct. Mater. 2008, 18, 726−731. (13) Park, S.; Kwon, O.-H.; Kim, S.; Park, S.; Choi, M.-G.; Cha, M.; Park, S. Y.; Jang, D.-J. J. Am. Chem. Soc. 2005, 127, 10070−10074. (14) Park, S.; Kwon, J. E.; Kim, S. H.; Seo, J.; Chung, K.; Park, S.-Y.; Jang, D.-J.; Medina, B. M.; Gierschner, J.; Park, S. Y. J. Am. Chem. Soc. 2009, 131, 14043−14049. (15) Mutai, T.; Ohkawa, T.; Shono, H.; Araki, K. J. Mater. Chem. C 2016, 4, 3599−3606. (16) Mutai, T.; Sawatani, H.; Shida, T.; Shono, H.; Araki, K. J. Org. Chem. 2013, 78, 2482−2489. (17) Mutai, T.; Tomoda, H.; Ohkawa, T.; Yabe, Y.; Araki, K. Angew. Chem., Int. Ed. 2008, 47, 9522−9524. (18) Shigemitsu, Y.; Mutai, T.; Houjou, H.; Araki, K. J. Phys. Chem. A 2012, 116, 12041−12048. (19) Skonieczny, K.; Ciuciu, A. I.; Nichols, E. M.; Hugues, V.; Blanchard-Desce, M.; Flamigni, L.; Gryko, D. T. J. Mater. Chem. 2012, 22, 20649−200664. (20) Ciuciu, A. I.; Flamigni, L.; Skonieczny, K.; Gryko, D. T. Phys. Chem. Chem. Phys. 2013, 15, 16907−16916. (21) Kim, S.; Seo, J.; Park, S. Y. J. Photochem. Photobiol., A 2007, 191, 19−24. (22) Tsai, H.-H. G.; Sun, H.-L. S.; Tan, C.-J. J. Phys. Chem. A 2010, 114, 4065−4079. (23) Chen, K.-Y.; Hsieh, C.-C.; Cheng, Y.-M.; Lai, C.-H.; Chou, P.-T. Chem. Commun. 2006, 4395−4397. (24) Piechowska, J.; Gryko, D. T. J. Org. Chem. 2011, 76, 10220− 10228. (25) Tseng, H.-W.; Lin, T.-C.; Chen, C.-L.; Lin, T.-C.; Chen, Y.-A.; Liu, J.-Q.; Hung, C.-H.; Chao, C.-M.; Liu, K.-M.; Chou, P.-T. Chem. Commun. 2015, 51, 16099−16102.

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.joc.7b01967. Spectroscopic data, theoretical calculations, and crystallographic data (PDF) Crystallographic data of 1 (CCDC 1563748) in cif format (CIF) Crystallographic data of 2 (CCDC 1563749) in cif format (CIF)

■

Article

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected] ORCID

Koji Takagi: 0000-0002-6107-0761 Takuya Nakashima: 0000-0002-5311-4146 Ryoichi Fukuda: 0000-0003-3001-4190 Masahiro Ehara: 0000-0002-2185-0077 Notes

The authors declare no competing financial interest. 12179

DOI: 10.1021/acs.joc.7b01967 J. Org. Chem. 2017, 82, 12173−12180

Article

The Journal of Organic Chemistry (26) Gryko, D. T.; Piechowska, J.; Galezowski, M. J. Org. Chem. 2010, 75, 1297−1300. (27) Takagi, K.; Ito, K.; Yamada, Y.; Nakashima, T.; Fukuda, R.; Ehara, M.; Takeuchi, D. Chem. Lett. 2017, 46, 1372−1375. (28) Hsu, Y.-H.; Chen, Y.-A.; Tseng, H.-W.; Zhang, Z.; Shen, J.-Y.; Chuang, W.-T.; Lin, T.-C.; Lee, C.-S.; Hung, W.-Y.; Hong, B.-C.; Liu, S.-H.; Chou, P.-T. J. Am. Chem. Soc. 2014, 136, 11805−11812. (29) Parada, G. A.; Markle, T. F.; Glover, S. D.; Hammarström, L.; Ott, S.; Zietz, B. Chem. - Eur. J. 2015, 21, 6362−6366. (30) Skonieczny, K.; Yoo, J.; Larsen, J. M.; Espinoza, E. M.; Barbasiewicz, M.; Vullev, V. I.; Lee, C.-H.; Gryko, D. T. Chem. - Eur. J. 2016, 22, 7485−7496. (31) Chan, B. K.; Ciufolini, M. A. J. Org. Chem. 2007, 72, 8489− 8495. (32) Skonieczny, K.; Gryko, D. T. J. Org. Chem. 2015, 80, 5753− 5763. (33) Wang, H.; Zhang, H.; Abou-Zied, O. K.; Yu, C.; Romesberg, F. E.; Glasbeek, M. Chem. Phys. Lett. 2003, 367, 599−608. (34) Chou, P.-T.; Chen, Y.-C.; Yu, W.-S.; Chou, Y.-H.; Wei, C.-Y.; Cheng, Y.-M. J. Phys. Chem. A 2001, 105, 1731−1740. (35) Chou, P.-T.; Yu, W.-S.; Cheng, Y.-M.; Pu, S.-C.; Yu, Y.-C.; Lin, Y.-C.; Huang, C.-H.; Chen, C.-T. J. Phys. Chem. A 2004, 108, 6487− 6498. (36) Sakai, K.; Ishikawa, T.; Akutagawa, T. J. Mater. Chem. C 2013, 1, 7866−7871. (37) Zhang, Y.; Wang, J.-H.; Zheng, W.; Chen, T.; Tong, Q.-X.; Li, D. J. Mater. Chem. B 2014, 2, 4159−4166. (38) Mosquera, M.; Penedo, J. C.; Ríos Rodíguez, M. C.; RodríguezPrieto, F. J. Phys. Chem. 1996, 100, 5398−5407. (39) Gaussian 09, Revision E.01. See the Supporting Information for full details. (40) Adamo, C.; Barone, V. J. Chem. Phys. 1999, 110, 6158−6170. (41) Dunning, T. H., Jr. J. Chem. Phys. 1989, 90, 1007−1023. (42) Kendall, R. A.; Dunning, T. H., Jr; Harrison, R. J. J. Chem. Phys. 1992, 96, 6796−6806. (43) Woon, D. E.; Dunning, T. H., Jr. J. Chem. Phys. 1993, 98, 1358− 1371. (44) Cancès, E.; Mennucci, B.; Tomasi, J. J. Chem. Phys. 1997, 107, 3032−3041. (45) Mennucci, B.; Cancès, E.; Tomasi, J. J. Phys. Chem. B 1997, 101, 10506−10517. (46) Cancès, E.; Mennucci, B. J. Math. Chem. 1998, 23, 309−326. (47) Cossi, M.; Barone, V. J. Chem. Phys. 2001, 115, 4708−4708. (48) Cammi, R.; Corni, S.; Mennucci, B.; Tomasi, J. J. Chem. Phys. 2005, 122, 104513. (49) A short history of SHELX: Sheldrick, G. M. Acta Crystallogr., Sect. A: Found. Crystallogr. 2008, A64, 112−122.

12180

DOI: 10.1021/acs.joc.7b01967 J. Org. Chem. 2017, 82, 12173−12180