Synthesis and investigation on optical and electrochemical properties of 2,4-diaryl-9-chloro-5,6,7,8-tetrahydroacridines

A facile synthesis of 2,4-diaryl-9-chloro-5,6,7,8-tetrahydroacridine derivatives is reported which is based on POCl3-mediated cyclodehydration followed by double Suzuki–Miyaura cross-coupling. The absorption and fluorescence properties of the obtained products were investigated and their HOMO/LUMO energy levels were estimated by cyclic voltammetry measurements. Besides, density functional theory calculations were carried out for further exploration of their electronic properties.

While considerable attention has been devoted to the photophysical properties of acridines, not much work has been reported related to their partially hydrogenated analogues, even though they were used as electron-donor groups for OLED applications [42,43]. On the other hand, tetrahydroacridines have received much attention in medicinal chemistry, due to their ability to inhibit topoisomerase enzymes and block the DNA transcription [44,45]. In particular, they have been widely explored for the treatment of Alzheimer's disease [46][47][48][49][50][51], human cancer [52,53], and tuberculosis [54] (Figure 1).
Taking into consideration the significant medicinal potential of tetrahydroacridines and the lack of knowledge concerning their optoelectronic properties, the search for new candidates is of ongoing interest. Recently, considerable attention has been given to acridines in our laboratory and we developed new synthetic methods for dibenzoacridines and acridones based on Pd-catalyzed cross-coupling reactions [55,56]. More recently, we have reported the synthesis of alkynylated 5,6,7,8-tetrahydroacridines [57]. In continuation to our previous work and as a part of our interest in new organic materials [58][59][60], we report herein the synthesis of hitherto unknown 2,4-diaryl-9chloro-5,6,7,8-tetrahydroacridine derivatives and their optical and electrochemical properties investigation by the use of UV-vis and emission spectroscopy, CV measurements as well as DFT calculations.
Our primary screening consisted in evaluating the effect of the catalyst amount. Several attempts were performed showing that a low concentration of the catalyst was required to avoid the formation of elevated amounts of homocoupling and side products. We found that reducing the amount of Pd(PPh 3 ) 4 from 5 mol % to 1 mol % gave higher yields of 2,4-diphenyl-9chloro-5,6,7,8-tetrahydroacridine (4a) ( Table 1, entries 1-6). Towards a better understanding of the regioselectivity of the coupling, we carried out a series of experiments with decreased amounts of phenylboronic acid. In all cases, the cross-coupling took place at both bromine atoms of 4a and we could never isolate the corresponding monoarylated coupling product. The replacement of Pd(PPh 3 ) 4 by PdCl 2 (PPh 3 ) 2 afforded 4a in almost the same yield (Table 1, entries 5 and 7). Concerning the impact of the base, we found that the application of K 2 CO 3 or Cs 2 CO 3 instead of K 3 PO 4 did not lead to an improvement of the yield (Table 1, entries 8 and 9). However, increasing the amount of K 3 PO 4 to 4 equivalents gave a nearly quantitative yield (Table 1, entry 12). We believe that a supplementary activation of the boronic acid, by adding a high amount of base, improved the nucleophilicity and facilitated the transmetalation. The use of dioxane instead of toluene (Table 1, entry 11) gave again a very good yield (89%). However, the employment of THF resulted in a decreased yield (Table 1, entry 10). The best result for the Suzuki-Miyaura cross-coupling between 2 and 3a was obtained using 1 mol % of Pd(PPh 3 ) 4 and 4 equivalents of K 3 PO 4 in toluene at 100 °C for 4 hours.
With our optimized conditions in hand, we examined the scope of the reaction of 2 with other arylboronic acids 3b-g (Scheme 2).  As expected, we found that electron-donating groups located in the arylboronic acid improved the yield. The cross-coupling of 2 with arylboronic acids 3b and 3d, bearing a methoxy group in position 2 or 4, afforded the corresponding products 4b and 4d in excellent yields. Arylboronic acids containing an alkyl group in different positions afforded good yields for products 4e and 4g. With the substrate containing a trifluoromethoxy group (OCF 3 ), a longer reaction time of 6 hours was required to obtain the product 4c in decreased yield (57%).

Photophysical properties.
We studied the steady-state absorption and emission of selected tetrahydroacridines 4a-d to gain insights into their photophysical properties. The UV-vis absorption spectra were measured in diluted dichloromethane solutions (c = 1 × 10 −5 M) at room temperature. Table 2 summarizes the obtained spectral data, fluorescence quantum yields, stokes shifts, and optical band gap energies.
As shown in Figure 2, most of the compounds show a broad band at around 320-360 nm in the UV region assigned to the π-π* electronic transitions. The phenyl unsubstituted derivative 4a exhibited a main wide band with a shoulder peak at 329 nm. Derivatives 4b and 4d, bearing methoxy substituents, showed similar optical absorptions to those of 4a with slight red shifts.
In case of the trifluoromethoxy-substituted derivative 4c, a larger red shift of 23 nm was observed. The optical gap energies of tetrahydroacridines 4a-d were estimated from the onset point of the absorption spectra [65]. The parent derivative 4a shows an onset of absorption at 363 nm and its optical band gap was deduced to be around 3.37 eV. Derivatives 4b and 4d showed similar values of 3.20 eV and 3.18 eV, respectively. In contrast, compound 4c showed a decreased value of 3.09 eV.
The emission spectra of tetrahydroacridines 4a-d were measured with an excitation line at 325 nm. They were dominated by broad bands ranging from 365 to 500 nm ( Figure 3). The emission spectrum of 4a appeared as a blue emission at 393 nm. Trifluoromethoxy derivative 4c exhibited a red shift of 10 nm as compared to 4a. Methoxy derivatives 4b and 4d  showed similar emission bands at around 410 nm and gave larger red shifts. Their bands at lower energy may be attributed to the intramolecular charge transfer (ICT) from the electrondonating methoxy groups to the tetrahydroacridine core.
The Stokes shifts of 4a-d were calculated based on their absorption and emission spectra. Derivative 4b, containing an electron-donating methoxy substituent in para-position, gave the highest value of 5770 cm −1 . The fluorescence quantum yields of 4a-d were calculated with a comparative method, where quinine sulfate (SQ) in 0.1 M H 2 SO 4 was used as standard [66].
The fluorescence and absorbance spectra for quinine sulfate and product 4b are given in Figure 4 and Figure 5, respectively. In addition, their plots of fluorescence intensities against absorbances are shown in Figure 6A and 6B. The tetrahydroacridine derivative 4b gave the highest fluorescence quantum yield of 8.2%. In order to elucidate the experimental observations and to further study the electronic properties, DFT calculations were carried out on the fully ground state at the restricted B3LYPlevel with 6-31G(d) basis set using dichloromethane as a continuum solvent model. The optimized chemical structures of 4a-d are given in Figure 7 with selected geometrical parameters including dihedral angles and bonding and non-bonding distances (Table 3). It can be seen from Figure 7 that the four obtained structures are non-planar, due to steric hindrance and electronic repulsions. For compound 4a, considerable distortions of −38.61 and −49.60 degrees between the tetrahydroacridine plane and the two phenyl groups were observed, due to rotations around the C-2 and C-4 single bonds. These torsion angles did not change significantly for products 4b and 4c bearing two substituents located at the para-positions. However, when the substituents were methoxy groups located at the ortho-positions, the dihedral angles were larger with values of −45.66 and −58.22 degrees, presumably due to steric reasons. These geometries may affect the electronic parameters of tetrahydroacridines 4a-d and particularly their band gap energies.
On the other hand, the photophysical properties of products 4e-g were not significantly affected by the aryl substituent. Electronic effects of the alkyl group were, as expected, very weak and no significant additional distortions arise from a paraor ortho-ethyl group for products 4e and 4g. Therefore, they  gave the same absorption and emission as the parent product 4a.
As well, the phenyl ortho-substituent in the aryl ring for 4f was twisted out of the plane and was not involved in the electronic orbital distribution ( Figure 8).
As shown in Figure 9, the HOMO and LUMO frontier orbitals of products 4a-d are extended over the π-conjugated part of the molecules and, as expected, the cyclohexane ring is not involved. In all cases, the electron densities of HOMO were  localized in the π-bonding orbitals between the carbon backbone of the quinoline ring and its two external phenyls. The LUMO electron densities were mainly located in the π* antibonding orbitals outside the carbon cores. It was observed for compound 4c that the two trifluoromethoxy groups were twisted out of the plane and consequently did not participate in the orbital distribution. The HOMO electronic distributions of 4b and 4d were spread over the two non-twisted methoxy groups.
The calculated values of HOMO, LUMO, and the band gap energies of compounds 4a-d at the B3LYP level of theory with 6-31G(d) basis set using dichloromethane as a continuum solvent model are given in Figure 9. The energy gap values are affected by the type and the position of the substituent located at the phenyl groups. In fact, different values of 3.87 eV and 4.06 eV were observed for compounds 4b and 4d having a methoxy group at the ortho-or para-positions, respectively. The electron-withdrawing trifluoromethoxy groups in compound 4c lead to a decrease in the HOMO energy to −6.10 eV and LUMO energy to −1.86 eV giving an energy gap of 4.02 eV. The HOMO energy of derivative 4b was the highest and as a result it exhibited the lowest value of E gap and its electron conductivity was the best.
The molecular electrostatic potential maps for tetrahydroacridines 4a-d are shown in Figure 10. The red and blue colors visualize the region of the attractive and repulsive potential, respectively. For the phenyl-substituted derivative 4a, the bluecolored surface, located mainly at the cyclohexane ring, visualizes the electron deficiency (high electrostatic potential). The red regions, localized essentially at the nitrogen atom and at the two external phenyl groups, show the electron abundance (low electrostatic potential). The trifluoromethoxy group for 4c resulted in a decrease of electron density in the tetrahydroacridine core. In addition, this strong electron-withdrawing effect induced the appearance of blue surfaces around the phenyl groups leading to a significant decrease in their electronic densities. However, thanks to the π-donating effect of two methoxy groups for 4b and 4d, yellow-red regions are present in the phenyl groups and quinoline core.
The electrochemical behavior of compound 4c was studied by cyclic voltammetry (CV) in dry acetonitrile. The obtained voltammogram, recorded at a scanning rate of 50 mV/s, is given in Figure 11. The HOMO and LUMO energy levels were deduced from the onset potentials of the oxidation and reduction waves, respectively. The potential of the saturated Ag/AgCl reference electrode was calibrated using the ferrocene/ferrocenium (Fc/Fc+) redox system [67], the HOMO and LUMO energies were calculated by using the following equations: Here, V FOC is the potential value of Fc/Fc+ under the same experimental conditions (0.52 V) and 4.8 eV is the HOMO energy   value of Fc/Fc+ with respect to zero vacuum level. The onset potentials for the oxidation and reduction of 3c were 1.50 eV and −1.36 eV, respectively. Thus, their LUMO and the HOMO energies were found to be −2.92 eV and −5.78 eV, respectively, giving rise to an electrochemical band gap of 2.86 eV. We noticed a difference of 0.23 eV between the optical and the electrochemical band gap which is attributed to the interface barrier between the electrode and the tetrahydroacridine, as well as to the exciton binding energy [68,69]. On the other hand, the calculated HOMO energy was more negative than the experimental value, while the calculated LUMO energy was less negative. This can be explained by the fact that the calculations were achieved based on an equilibrium conformer.

Conclusion
In conclusion, we developed a convenient synthesis of novel 2,4-diaryl-9-chloro-5,6,7,8-tetrahydroacridines starting from commercially available anthranilic acid. Their photophysical properties show an impact of the aryl substituents on absorption and fluorescence. Methoxy substituents lead to a red shift of the emission and exhibited the highest fluorescence quantum yields. According to the theoretical studies, all products adopted a distorted molecular geometry. In contrast to methoxy groups, the trifluoromethoxy groups are twisted out of plane and, therefore, are not involved in the electronic orbital distribution. Experimental energy band gaps of the prepared tetrahydroacridines were found to be an average of 3 eV which makes them promising candidates for optoelectronic applications.

Materials and measurements
Anhydrous solvents and chemicals were purchased from Sigma-Aldrich and used without further purification. All reac-tions were carried out under an inert argon atmosphere and were monitored by thin-layer chromatography (TLC). Organic compounds were purified using commercial Merck silica gel (0.043-0.06 mm) with a fluorescence indicator (the visualization was performed under UV light at 254 nm). All solvents for work-up and column chromatography were distilled before use. NMR data were recorded in CDCl 3 (tetramethylsilane as the internal standard) on Bruker ARX 300 instruments. Signals due to the solvent are: CHCl 3 : δ 7.26 for 1 H and δ 77.16 for 13 C. Peak characterization of 1 H NMR spectra: s = singlet, d = doublet, t = triplet, q = quartet, m = multiplet. Chemical shifts were given in ppm (δ) relative to tetramethylsilane (SiMe 4 ) as internal standard. High-resolution MS (HRMS-ESI) was performed on an Agilent 1969 A TOF. The photophysical studies were carried out in freshly prepared dichloromethane solutions with concentrations of 1 × 10 −5 M. The UV-vis spectra were recorded on a Shimadzu 2401 PC spectrophotometer in quartz cuvettes with a path length of 1 cm. Emission spectra were recorded on a Perkin-Elmer LS 50B spectrofluorimeter. Cyclic voltammetry was performed in anhydrous acetonitrile solution containing 0.1 M of tetrabutylammonium tetrafluoroborate (n-Bu 4 NBF 4 ) as a supporting electrolyte by using a three-electrode cell with a glassy carbon electrode as the working electrode, a platinum wire as the counter electrode, and an Ag/AgCl electrode as the reference electrode. All the measurements were scanned in negative and positive directions at a scan rate of 50 mV/s, as controlled by a PGSTAT30 Autolab potentiostat.

Theoretical calculations
The theoretical studies were realized in vacuum with the Gaussian 09 program [70]. The geometry of the equilibrium conformer at the ground state was first found at AM1 level. Then, further optimization through density functional theory (DFT) approach [71] at the restricted Becke, 3-parameter, Lee-Yang-Parr hybrid functional (B3LYP) with standard basis set 6-31G (d) were carried out.

Supporting Information
Supporting Information File 1 Synthesis and analytical data of starting compound 2 and copies of spectra for the synthesized compounds.

Funding
This study was supported by the Ministry of Higher Education and Scientific Research-Tunisia. Financial support from the DAAD (scholarship for N. T.) is gratefully acknowledged.