Balancing charge-transfer strength and triplet states for deep-blue thermally activated delayed fluorescence with an unconventional electron rich dibenzothiophene acceptor

Anal. Calc. for C 56 H 32 N 6 : C, 84.62; H, 8.08; N, 3.40. Found: C, 84.48; H, 8.02; N, 3.27. m.p. = 255-257  C. –pump-thaw cycles. The films in zeonex matrix were fabricated by drop-casting onto quartz substrates. Steady-state absorption and emission spectra were acquired using a UV-3600 Shimadzu spectrophotometer and a Jobin Yvon Horiba Fluoromax 3, respectively. PLQYs were measured using a calibrated Horiba Quanta-Φ integrating sphere coupled to a Fluorolog3 spectrometer using 315 nm excitation. Data was processed using the included FluorEssence software.


S1a. General experimental details
All reactions were carried out under an argon atmosphere unless otherwise stated. Starting materials were purchased commercially and were used as received. Solvents were dried using an Innovative Technology solvent purification system and were stored in ampoules under argon. TLC analysis was carried out using Merck Silica gel 60 F254 TLC plates and spots were visualised using a TLC lamp emitting at 365, or 254 nm. Silica gel column chromatography was performed using silica gel 60 purchased from Sigma Aldrich. 1 H and 13 C{ 1 H} NMR spectroscopy was carried out on Bruker AV400, Varian VNMRS 500 and 700, and Varian Inova 500 NMR spectrometers. Residual solvent peaks were referenced as described in the literature, 1 and all NMR data was processed in MestReNova V12. Melting points were carried out on a Stuart SMP40 machine with a ramping rate of 4 °C min −1 . Videos were replayed manually to accurately determine the melting point. High resolution mass spectrometry was carried out on a Waters LCT Premier XE using ASAP ionization and TOF detection. Samples were analyzed directly as solids using N 2 at 350 °C. Elemental analysis was performed on an Exeter Analytical E-440 machine. TGA analysis was carried out on a Perkin Elmer Pyris 1 machine under nitrogen gas at 20 mL min −1 . Measurements were carried out from 25 °C -500 °C at 10 °C min −1 .

2,8-Bis(9,9-dimethylacridin-10(9H)-yl)dibenzo[b,d]thiophene (DAc-DBT)
2,8-Dibromodibenzothiophene (268 mg, 0.78 mmol, 1 eq.) and 9,9-dimethyl-9,10dihydroacridine (327 mg, 1.57 mmol, 2 eq.) were dried under vacuum for 30 min in a twoneck 100 mL round-bottomed flask fitted with a reflux condenser. The flask was back-filled with argon and dry toluene (20 mL) was added via syringe. The reaction mixture was bubbled with argon for 30 min, then Pd 2 (dba) 3 ·CHCl 3 (41 mg, 39 μmol, 0.05 eq.) and HP t Bu 3 BF 4 (23 mg, 79 μmol, 0.10 eq.) were added and the reaction mixture was bubbled with argon for a further 30 min. NaO t Bu (225 mg, 2.34 mmol, 3 eq.) was added under a high flow of argon and the reaction was then heated to 115 C (DrySyn kit temperature) with stirring for 17 h. At 3 the end of the reaction CH 2 Cl 2 (70 mL) was added followed by water (70 mL). The organic layer was separated and the aqueous layer was extracted further with CH 2 Cl 2 (2 × 70 mL). The organic extracts were combined, and were dried with MgSO 4 and filtered. The solvent was removed under reduced pressure and the crude mixture was purified by silica gel column chromatography with gradient elution from 20% CH 2 Cl 2 /hexane switching to 25% CH 2 Cl 2 . Removal of solvent under reduced pressure gave a yellow solid. Recrystallization by addition of hexane to a concentrated solution of product in CH 2 Cl 2 gave DAc-DBT as a white crystalline solid (205 mg, 45%). Crystals suitable for X-ray diffraction were obtained by slow evaporation from CD 2 Cl 2 and allowing for solvent evaporation. 1

2,8-Bis(2,7-di(tert-butyl)-9,9-dimethylacridin-10(9H)-yl)dibenzo[b,d]thiophene (DtBuAc-DBT)
2,8-Dibromodibenzothiophene (0.17 g, 0.77 mmol, 1 eq.) and compound 1 (0.50 g, 1.55 mmol, 2.2 eq.) were dried under vacuum for 30 min in a two-neck round-bottomed 100 mL flask fitted with a reflux condenser. The flask was back-filled with argon for 30 min, then Pd 2 (dba) 3 •CHCl 3 (0.04 g, 0.038 mmol, 0.05 eq.) and HP t Bu 3 BF 4 (22 mg, 0.07 mmol, 0.1 eq.) were added and the reaction mixture was bubbled with argon for 30 min. t BuONa (0.22 g, 2.33 mmol, 3 eq.) was added under a high flow of argon and the reaction was heated then to 115C (DrySyn kit temperature) with stirring for 17 h. After being cooled to room temperature, the reaction mixture was extracted with CH 2 Cl 2 (3 × 100 mL). The organic layer was dried with MgSO 4 and filtered. The solvent was removed under reduced pressure and the crude mixture was purified by silica gel chromatography with gradient elution from 40% v/v CH 2 Cl 2 /hexane switching to 100% CH 2 Cl 2 in 20% increasing increments. Removal of solvent under reduced pressure resulted in product as a white solid. Recrystallization from boiling hexane with dropwise addition of CH 2 Cl 2 followed by cooling to 0 C gave pure product as a white crystalline solid (0.32 g, 51%). Crystals suitable for X-ray diffraction were obtained by slow evaporation from CH 2 Cl 2 / hexane mixture of solvents. 1  9,9-Dimethyl-9,10-dihydroacridine was prepared according to the literature procedure. 2

S2. X-Ray Crystallography
X-ray single crystal data were collected on a Bruker 3-circle D8 Venture diffractometer with a Photon100 CMOS detector, using Mo-K radiation ( = 0.71073 Å) from a IμSmicrosource with focusing mirrors. The crystals were cooled with a Cryostream (Oxford Cryosystems) open-flow N 2 cryostat. The structures were solved by direct methods using SHELXS program, 4 and refined using SHELXL software 5 on OLEX2 platform. 6 Selected crystal data and experimental details are listed in Table S1, full crystallographic information (including structure factors) in CIF format has been deposited with Cambridge Crystallographic Data Centre.
DAc-DBT undergoes a reversible phase transition on cooling below 170 K, whereupon the structure is modulated with doubling of the a parameter. The room-temperature phase was characterized at 200 K and the low-temperature one at 120 K. The structure of DtBuAc-DBT was also determined at 120 and 200 K, but was found to be essentially the same at both temperatures.

S3. Computational Details
Computations were performed with the Gaussian 09 package 7 using various density functional theory (DFT) methods. The isolated molecules were optimized at the rBMK/6-31G(d) level in gas phase. The spectroscopic properties of the molecules were calculated by means of time-dependent DFT (TD-DFT) 8 calculations employing 6-31G(d) basis set. The BMK functional was preferred for the geometry optimization and the prediction of the optical properties of all the studied molecules as it was shown to be pertinent for the description of the low energy band in the absorption of D-A CT compounds. 9-11 Multiwfn 12 software was used to evaluate molecular fragment contributions to occupied and virtual orbitals.

S4. Optical Characterization
Two types of samples were studied in this work: solutions (10 −3 to 10 −5 M) and films produced in zeonex matrix (1% w/w). All of the solutions were diluted in different solvents (ethanol, 1,2-dichlorobenzene, toluene and methylcyclohexane) and stirred for up to 24 hours. To remove oxygen for emission measurements, solutions were degassed by four freeze -pump-thaw cycles. The films in zeonex matrix were fabricated by drop-casting onto quartz substrates. Steady-state absorption and emission spectra were acquired using a UV-3600 Shimadzu spectrophotometer and a Jobin Yvon Horiba Fluoromax 3, respectively. PLQYs were measured using a calibrated Horiba Quanta-Φ integrating sphere coupled to a Fluorolog3 spectrometer using 315 nm excitation. Data was processed using the included FluorEssence software.
Prompt fluorescence (PF), delayed fluorescence (DF) and phosphorescence spectra and timeresolved decays were recorded using nanosecond gated luminescence and lifetime measurements (from 800 ps to 1 s) with either a high energy pulsed Nd:YAG laser emitting at 355 nm (EKSPLA) and a N 2 laser emitting at 337 nm with pulse width 170 ps. Emission was focused onto a spectrograph equipped with 300 lines/mm grating of 500 nm or 1000 nm base wavelength and detected on a sensitive gated iCCD camera (Stanford Computer Optics) with sub-nanosecond resolution. Time-resolved measurements were performed by exponentially increasing the gate and delay times. The delay and integration times are chosen in a way that the next delay is set at a time longer than the previous delay+integration time. Therefore, no spectral overlap exists between the spectra corresponding to successive delays. The curve obtained directly from this process does not represent the real luminescence decay. However, this is easily corrected by integrating the measured spectra and dividing the integral by the corresponding integration time. In this way, each experimental point represents a snap-shot of the number of photons emitted per second at a time t = delay + (integration time)/2. The luminescence decay is then obtained by plotting each experimental point versus time, and fitting with sum of exponentials when required. In this way we are able to collect the entire emission spectrum decaying over 8 decades in a single experiment 13 . For initial development of these methods see previously published literature. 14

S5. Cyclic Voltammetry Measurements
Electrochemical measurements were performed in solutions of 0.1 M Bu 4 NBF 4 (99%, Sigma Aldrich, dried) in dichloromethane (DCM, 99.9%, Extra Dry, stabilized, AcroSeal®, Acros Organics) at room temperature. Solutions were prepared with 1.0 mM concentrations of the D-A-D compound and purged with nitrogen prior to measurement. The electrochemical cell is composed of three electrodes: a platinum disc with 1 mm diameter of working area as a working electrode, an Ag/AgCl electrode as a reference electrode and a platinum wire as an auxiliary electrode. All cyclic voltammetry (CV) measurements were performed at room temperature with a potential scan rate of 50 mV/s and calibrated against a ferrocene/ferrocenium (Fc/Fc + ) redox couple.
The onset potential was determined from the intersection of two tangents drawn at the rising and background current of the CV. The ionization potential (IP) was calculated from the oxidation (E ox ) potential, using the following equation: IP = E ox + 5.1. 15 HOMO energy levels were determined using CV analysis by the estimation of IPs which are similar to the HOMO energies. 16 As the reduction potential is out of range even in DMF solution, LUMO energies were calculated according to the optical band gap E g , which was determined from the onset of the UV-Vis absorption band in DCM.  For DAc-DBT, a new peak is observed on the reverse scan at a lower potential resulting from degradation or reaction of the molecule leading to the formation of a conducting layer on the working electrode. This peak was not observed in the substituted derivatives DtBuAc-DBT and DOMeAc-DBT suggesting that its origin may be reactivity at the 2,7-positions of acridine. Similar behaviour was observed in other unsubstituted acridine derivatives 17 pointing to the lack of electrochemical stability. Furthermore, the HOMO energies are elevated with the increasing donor strength, which is consistent with the calculations shown in Figure S4. LUMO energies are determined from the HOMO levels and the corresponding optical band gaps. The HOMO and LUMO energy levels are shown in Table S3.

S8. Devices
Organic light-emitting diodes (OLEDs) were fabricated on patterned indium-tin-oxide (ITO) coated glass (VisionTek Systems) with a sheet resisitance of 15 Ω/cm 2 using vacuum thermal evaporation. The substrates were sonicated for 15 minutes each in acetone and then IPA. After oxygen-plasma cleaning, the substrates were loaded into a Kurt J. Lesker Super Spectros 200 deposition chamber. All organic and cathode layers were thermally evaporated at a pressure below 10 -7 mbar, at evaporation rates in the range of 0.1-0. The evaporated devices were encapsulated under inert atmosphere using UV-curable epoxy (DELO Katiobond) along the outer edges of the active emitting area with a glass coverslip. Characterization of the OLED devices was conducted in a 10-inch integrating sphere (Labsphere) coupled with a calibrated fibre spectrometer (Ocean Optics USB4000) and connected to a Keithley 2400 source measure unit.