Crystallisation-enhanced bulk hole mobility in phenothiazine-based organic semiconductors

A series of three novel donor-acceptor systems based on C(3)-malononitrile-substituted phenothiazines was synthesised in good overall yields and their thermal, spectroscopic, and electrochemical properties were characterised. The compounds were prepared through a sequence of Ullmann-coupling, Vilsmeier-Haack formylation and Knoevenagel-condensation, followed by Suzuki-coupling reactions for introduction of aryl substitutents at C(7) position of the phenothiazine. The introduction of a donor unit at the C(7) position exhibited a weak impact on the optical and electrochemical characteristics of the compounds and led to amorphous films with bulk hole mobilities in the typical range reported for phenothiazines, despite the higher charge delocalisation as attested by computational studies. In contrast, highly ordered films were formed when using the C(7)-unsubstituted 3-malononitrile phenothiazine, exhibiting an outstanding mobility of 1 × 10−3 cm2 V−1 s−1, the highest reported for this class of compounds. Computational conformational analysis of the new phenothizanes suggested that free rotation of the substitutents at the C(7) position suppresses the ordering of the system, thereby hampering suitable packing of the new materials needed for high charge carrier mobility.


Details on synthesis
Synthesis of 2-bromo-6-butoxynaphthalene (6). 6-bromo-2-napthol (4 g, 18 mmol), potassium hydroxide (1.2 g, 21 mmol) and dimethyl sulfoxide (50 mL) were added into a 250 mL two necked round bottom flask, and stirred for 30 min. Then, 1-bromo butane (2.3 mL, 21 mmol) was added drop wise to the mixture and reaction mixture was stirred overnight at room temperature. After completion of the reaction, the mixture was poured into water (200 mL) and extracted with ethyl acetate (3× 50 mL). The ethyl acetate solution was washed with brine solution and water and dried over anhydrous sodium sulphate, filtered, and finally ethyl acetate was evaporated with a rotary evaporator. The crude product was purified by silica gel column chromatography using petroleum ether as the eluent to give title compound (4.7 g, 95 % yield) as a white solid. 1   phosphorous oxychloride (3.65 mL) was added drop wise at 0 °C. The reaction mixture was stirred at room temperature for 30 min, and then the solution of 1 (4g, 13 mmol) in 1,2dichloroethane (30 mL) was added and the mixture was heated to 80 o C for 12 h. The reaction mixture was poured into the water (100 mL), extracted with ethyl acetate (3x50 mL), and washed with brine solution and water. The ethyl acetate solution was dried over anhydrous sodium sulphate and filtered. Ethyl acetate was evaporated on rotary evaporator. The crude product was purified by silica gel column chromatography using petroleum ether: dichloromethane (7:3 v/v) as an eluent to give the title compound (3.7 g, 85 % yield) as a yellow solid. 1

Synthesis of 7-bromo-10-(4-methoxyphenyl)-10H-phenothiazine-3-carbaldehyde (3). 10-(4-
Methoxyphenyl)-10H-phenothiazine-3-carbaldehyde (3 g, 8 mmol) was dissolved into a two necked round bottom flask in chloroform (50 mL). After cooling the mixture to 0 o C for 30 min, previously prepared mixture of N-bromosuccinimide (1.92 g, 10.7 mmol) in chloroform (15 mL) was added drop wise into the flask. Then the reaction mixture was stirred overnight at room temperature. After completion of reaction, the reaction mixture was poured into water, extracted with chloroform (3×50 mL), and washed with brine solution and water. The chloroform solution was dried over sodium sulphate, filtered and chloroform was evaporated on a rotary evaporator. The crude product was purified on silica gel chromatography using petroleum ether: ethyl acetate (9:1, v/v) as an eluent to afford title compound (3.75 g, 85% yield) as a pale saffron solid. 1

Computational Details
All calculations were carried out using the Gaussian09 1 software package without symmetry constrains. Solvent effects (chloroform) were considered in every calculation using the Polarizable Continuum Model (PCM) initially devised by Tomasi and coworkers 2-4 as implemented on Gaussian 09, with radii and non-electrostatic terms for Truhlar and coworkers' SMD solvation model. 5 Density Functional Theory (DFT) and Time-Dependent DFT (TD-DFT) were used for computation of the ground and excited state properties of the phenothiazines, respectively. All calculations have been performed using the PBE1PBE functional and 6-31G(d, p) 6-10 basis set. That functional uses a hybrid generalised gradient approximation (GGA), including 25% mixture of Hartree-Fock 11 exchange with DFT 12 exchange-correlation, given by Perdew, Burke and Ernzerhof functional (PBE). 13,14 Extensive TD-DFT calculations studies with different functionals have demonstrated that global hybrid methods containing between 22-25% of exact exchange provide better match with reference data. 15 Vertical excitation calculations of compound O-1 with PBE1PBE, CAM-B3LYP and B3LYP functionals and 6-31G(d, p) basis set using PCM were compared with the experimental UV spectrum in chloroform. PBE1PBE was adopted for the computational study since the λmax calculated (494 nm) was in considerable better agreement with the experimental value (490 nm) than the one obtained with CAM-B3LYP (400 nm) or B3LYP (516 nm). Successful uses of PBE1PBE in the studies of vertical excitation systems have been reported for aromatic [16][17][18] and push-pull systems. 19,20 S-20 The computational data described was obtained by using the following computational protocol: 1) The ground state geometry of each molecule has been fully optimized with default cut-offs on forces and step size to determine convergence.
2) The analytical calculation of the vibrational frequencies at the same level of theory verified the optimized structure by checking that they corresponded to true minima of the potential energy surface by the absence of imaginary frequencies.
3) The first six low-lying excited states have been determined within the vertical TD-DFT, with the default non-equilibrium solvation procedure. The less energetic allowed transition (S0→S1*) were mainly composed by HOMO→LUMO transitions. 4) The minimum energy point on the excited state potential energy surface was calculated by a TD-DFT geometry optimization of the first excited state in solution, with tight convergence criteria, with equilibrium, linear response solvation. 5) A Natural Population Analysis (NPA) [21][22][23][24][25][26][27][28] was performed as implemented on Gaussian 09 to study the electronic structure of the optimised species.

Computational analysis
The electronic nature of the newly synthesised phenothiazines O-1 and O-2 was explored by Density Functional Theory methods, as well as O-3 simplified by replacement of the nbutyl chain with a methyl (O-3(Me)). In the ground state the geometries of the phenothiazines are slightly bent in the phenothiazine heterocyclic ring and adopt a butterfly shape. The presence of different substituents at the C(7) position of the phenothiazine has little or no effect on the dihedral angles made by the two aromatic rings, being about 19-20° for the three molecules studied. Although alkyl N-substituted phenothiazines present larger dihedral angles (ca. 41°), the presence of the methoxyphenyl N-substituent induces the heterocyclic moiety to adopt a geometry closer to planarity. 29 On the other hand, the methoxyphenyl N-substituent preferentially adopts a perpendicular position to the unsubstituted phenyl ring in O-1 (Φ1) and its co-planarity demands for 7 kcal/mol (Fig.  S16). Regarding rotation of substituents at C(7) position, these adopt a non-planar geometry having a dihedral angle (Φ2) of ca. 34° with the phenothiazine ring system in O-2 and O-3(Me) (Fig. S16). Such slight distortion does not seem to encumber the resonance between the substituents and the phenothiazine.
S-21 Figure S16. Plane figures of phenothiazines optimized structures, dihedral angles and energy profiles for substituents rotation The electron density distribution in the LUMO of the phenothiazines is mainly localized on the malononitrile end group and the adjacent phenyl ring, while in O-2 and O-3(Me) HOMO orbitals are delocalized over the phenothiazine moiety and its C(7) substituents, as shown in Fig. 7. Clearly, introduction of electron rich substituents in the C(7) position of the phenothiazine increases their HOMO energy while keeping the energy of LUMO orbitals localized on the acceptor end. The methoxyphenyl N-substituent is not involved in the electron density of the frontier orbitals, as would be expected due to its perpendicularity with the phenothiazine system. The strongest transitions (S0→S1) are visibly dominated by HOMO→LUMO transitions for the three studied molecules, as determined by TDDFT (Table  S1). In the case of phenothiazine O-3(Me) a small contribution from HOMO-1→LUMO transition was found. The λmax found reflects the observed experimental trends, although the absolute values for the excitation energies are overestimated by the method (by ca. 0.9-1.1 eV). Further inspection into the S1 state and the distribution of natural charges in S0 compared to S1 (Fig. S17) confirms that the three systems behave similarly concerning the acceptor moiety and the methoxyphenyl N-substituent. The acceptor moiety is electronically enriched by 0.35 upon excitation, while the N-substituent barely changes its natural charge. On the other hand, while most of the charge transfer occurs from the phenothiazine ring into the malononitrile, the electron rich substituent at C(7) contributes to some extent to the internal charge transfer. In such instance the 6-methoxynaphtalene of O-3(Me) is slightly more electron donating than the 4-methoxyphenyl substituent in O-2, as demonstrated by the changes in natural charges (0.09 for O-2 and 0.12 for O-3(Me)). Figure S17. Differences in Natural Charges between ground state (S0) and excited state (S1), estimated by a time dependent DFT/PBE1PBE model.