Achieving Deep-Blue Thermally Activated Delayed Fluorescence in Nondoped Organic Light-Emitting Diodes through a Spiro-Blocking Strategy.

A deep-blue thermally activated delayed fluorescence (TADF) emitter TXADO-spiro-DMACF has been reported for nondoped organic light-emitting diodes (OLEDs) by integrating an appropriate blocking unit with the donor (D)-acceptor (A)-donor (D)-type TADF emitter via a spiro linkage. Benefiting from the characteristic perpendicular arrangement, the intermolecular interactions are expected to be weakened to some degree. As a result, TXADO-spiro-DMACF shows a very small bathochromic shift of 8 nm associated with a narrowed full width at half maximum of 54 nm on going from solution to the film. The corresponding nondoped device successfully achieves a bright deep-blue emission, revealing Commission Internationale de l'Eclairage coordinates of (0.16, 0.09) and a peak external quantum efficiency of 5.3% (5.3 cd/A, 5.9 lm/W). The results clearly indicate that spiro-blocking is a promising strategy to develop deep-blue TADF emitters capable of nondoped OLEDs.

Introduction

Pure organic molecules showing thermally activated delayed fluorescence (TADF) have emerged as the third generation electroluminescence (EL) materials for organic light-emitting diodes (OLEDs) nowadays.[1−3] Different from the conventional fluorescent emitters with a limited 25% internal quantum efficiency (IQE),[4,5] TADF molecules can further utilize 75% of electrically generated triplet excitons and thus realize a nearly unity IQE through efficient up-conversion from the lowest triplet state (T1) to the lowest singlet state (S1).[6,7] Meanwhile, they do not contain any noble metals but can achieve parallel device performance to phosphorescent complexes[8,9] in a much more different and cheaper way.[10−12] Promoted by these features, numerous TADF emitters that reveal different emission colors in the whole visible region have already been designed and used successfully to fabricate TADF-based OLEDs.[13−15]

Among them, deep-blue TADF emitters are of great importance, whose Commission Internationale de l’Eclairage (CIE) coordinate of y (CIE) ≤ 0.10 is required to increase the color gamut and reduce power consumption in display applications.[16,17] Furthermore, they are also indispensable for the construction of blue hyperfluorescence systems, where they act as a host matrix or an assistant dopant for the conventional fluorescent emitters so as to harvest both singlet and triplet excitons.[18,19] Early in 2012, Adachi et al. observed the first deep-blue TADF from a carbazole/sulfone derivative and obtained a maximum external quantum efficiency (EQE) of 9.9% and CIE coordinates of (0.15, 0.07).[20] After that, much effort has been paid to the design of efficient deep-blue TADF emitters,[21−24] and the EQE is improved to 19.2% together with CIE coordinates of (0.15, 0.10).[22] We note that all the currently reported progresses are made based on a doped device configuration. In this case, the host investigation and selection is still an ongoing activity especially for deep-blue TADF emitters because the requirement of a wide band gap as well as high triplet energy inevitably leads to unfavored carrier injection and transporting. As an alternative, nondoped OLEDs (also denoted as host-free OLEDs) are able to solve the problem, in which the emitting layer is composed of an emitter independently without any additional host.[25−28] However, the realization of deep-blue TADF in nondoped OLEDs remains a big challenge because most of deep-blue TADF emitters often suffer from a considerable spectral red-shifting,[29,30] and aggregation caused quenching[31,32] in their neat films.

On the other hand, a spiro geometry is beneficial to reduce the intermolecular interactions and thus suppress the spectral red-shifting and aggregation-caused quenching.[33,34] However, none of the spiro-based TADF molecules are reported to be suitable for nondoped devices with deep-blue emissions now.[35−39] Herein, we newly propose a spiro-blocking strategy toward deep-blue TADF emitters capable of nondoped OLEDs. As depicted in Figure , an appropriate blocking unit is integrated with the linear donor (D)–acceptor (A)–donor (D) type TADF emitter via a spiro linkage. Benefiting from the characteristic perpendicular arrangement, the intermolecular interactions would be weakened to some degree, and the realization of deep-blue TADF in nondoped OLEDs is within our expectations. As a proof of concept, a model compound TXADO-spiro-DMACF has been demonstrated, whose nondoped device gives a deep-blue EL with CIE coordinates of (0.16, 0.09) as well as a promising EQE of 5.3% (5.3 cd/A, 5.9 lm/W).

Figure 1

Spiro-blocking strategy to realize deep-blue TADF emitters for nondoped OLEDs.

Results and Discussion

Molecular Design

As for TXADO-spiro-DMACF (Figure ), 9,9-dimethylacridine and fluorene are used as the donor and acceptor, respectively, to constitute a linear D–A–D emitter (DMACF), and thioxanthene dioxide (TXADO) is chosen as the blocking unit, whose highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) energy levels play a key role on the formation of the spiro-blocking. In other words, the HOMO of TXADO should be lower than that of DMACF (Table S1) to ensure the HOMO distribution mainly on 9,9-dimethylacridine. At the same time, the LUMO of TXADO is higher than that of DMACF so that the LUMO distribution could be effectively limited on fluorene rather than TXADO. Combined with the large torsion angle between TXADO and DMACF (81.9°), the observations clearly indicate that the spiro-linked TXADO does not contribute to the charge transfer (CT) nature of the whole molecule, but acts as the blocking unit to prevent the intermolecular interactions. On the other hand, after the introduction of TXADO, the LUMO is greatly down from −1.30 eV of DMACF to −1.56 eV of TXADO-spiro-DMACF, while their HOMO levels remain nearly unchanged (Table S1). Thereby, S1 of TXADO-spiro-DMACF, estimated by the time-dependent density functional theory (TD-DFT) based on its optimized ground state geometry, is correspondingly reduced to 2.9937 eV compared with DMACF (3.2440 eV, Table S2). Given the small variation of T1 (here 3LE is the lowest triplet state) between DMACF and TXADO-spiro-DMACF, the energy difference between S1 and T1 (ΔEST) of TXADO-spiro-DMACF is predicted to be 0.1730 eV in the toluene solution, much lower than that of DMACF (0.3876 eV). In brief, the introduced TXADO can not only function as a blocking group but also lead to the decreased ΔEST to favor the efficient up-conversion from T1 to S1 by absorbing environmental thermal energy. This implies that TXADO-spiro-DMACF other than DMACF may be a good TADF emitter suitable for the nondoped device.

Figure 2

(a) Molecular structure; (b) molecular orbital distributions of HOMO and LUMO (the energy diagram is obtained by TD-DFT); and (c) crystal structure of TXADO-spiro-DMACF (marked with torsion angles).

Synthesis and Characterization

The synthetic route for TXADO-spiro-DMACF is presented in Scheme . Starting from 4,4′-dibromo-1,1′-biphenyl, the key intermediate 2,7-dibromospiro[fluorene-9,9′-thioxanthene]-10′,10′-dioxide (5) was firstly prepared through nitration, reduction, iodination, Kumada reaction, and cyclization.[4] Then, a C–N coupling between 5 and 9,9-dimethylacridine was performed to afford TXADO-spiro-DMACF with a yield of 96%. The final compound was fully confirmed by 1H and 13C NMR, MALDI-TOF mass spectrometry and elemental analysis (Figure S1). The decomposition temperature (Td, corresponding to a 5% weight loss) of TXADO-spiro-DMACF is determined to be above 460 °C, and no obvious glass transition is detected in the range of 35–300 °C (Figure S2). The rigid geometry of TXADO-spiro-DMACF should be responsible for the observed excellent thermal stability, which makes it favorable for vacuum deposition and long-term OLEDs. Moreover, according to the cyclic voltammetry (CV) and optical bandgap, the HOMO and LUMO levels are determined to be −5.33 and −2.23 eV, respectively (Figure S3 and Table ).

Scheme 1

Synthetic Route for TXADO-Spiro-DMACF

Reagents and conditions: (i) fuming HNO3, CH3COOH, 100 °C; (ii) Fe, CH3COOH, H2O, 100 °C; (iii) NaNO2, KI, HCl, 0 °C → 50 °C; (iv) Mg, I2, LiCl, CH3CH2Br, THF, 60 °C; (v) CH3COOH, HCl, H2O, 100 °C; (vi) Pd2(dba)3, P(t-Bu)3HBF4, t-BuONa, trimethylbenzene, 135 °C.

Table 1

Physical Properties of TXADO-Spiro-DMACF

	TXADO-spiro-DMACF
λabs [nm]a	284 (1.03 × 105), 364 (8.95 × 103)
λPL [nm]a	437 (62)
λPL [nm]b	445 (54)
τp [ns]c	3.8
τd [μs]d	106
ΦPLe	0.42 (0.41)
S1/T1/ΔEST [eV]f	3.14/2.86/0.28 (2.99/2.82/0.17)
Eg [eV]g	3.10
HOMO/LUMO [eV]h	–5.33/–2.23
Td [°C]i	464

Measured in toluene, the molar extinction coefficient and fwhm is given in the parentheses, respectively.

Measured in the neat film and fwhm is given in the parentheses.

Measured in the neat film under nitrogen in the range of 0–100 ns.

Measured in the neat film under nitrogen in the millisecond scale.

Measured in the neat film under nitrogen, and the value for 10 wt % doped film in TCTA is given in the parentheses.

S1 and T1 are determined from the onset of fluorescence and phosphorescence spectra in toluene, respectively. ΔEST is the energy difference between S1 and T1. The computational values are also listed in the parentheses for comparison.

Estimated from the absorption onset.

HOMO = −e[Eonset,ox + 4.8] V, LUMO = HOMO + Eg, where Eonset,ox is the onset value of the first oxidation potential.

Decomposition temperature corresponding to a 5% weight loss.

Photophysical Properties

The steady-state UV–vis absorption and photoluminescence (PL) spectra in toluene solutions for TXADO-spiro-DMACF, compared with its film PL spectrum, are shown in Figure a. Besides the intense absorption band below 330 nm, there is another weak but the discernible band in the range of 330–400 nm. As discussed above, this absorption could be reasonably attributed to the CT band from 9,9-dimethylacridine to fluorene. Also, a strong PL peaked at about 437 nm is observed for TXADO-spiro-DMACF in toluene, corresponding to an S1 of 3.14 eV. On going from solution to the film, the emission moves from 437 nm to a longer wavelength of 445 nm, and the full width at half maximum (fwhm) is slightly down from 62 to 54 nm. The very small red-shift (8 nm) as well as reduced fwhm of TXADO-spiro-DMACF implies that the intermolecular interactions are substantially inhibited in solid states because of spiro-blocking. This is further verified by the PL quantum yield (PLQY = 0.42) of TXADO-spiro-DMACF in the nondoped film, which is close to that of the 10 wt % doped film into the host tris(4-(9H-carbazol-9-yl)phenyl)amine (TCTA). Also, these results are in well agreement with the crystal packing in TXADO-spiro-DMACF (Figure S4), where TXADO is found to act as an efficient blocking unit between acceptor and acceptor, donor and donor, and acceptor and donor.

Figure 3

Photophysical properties of TXADO-spiro-DMACF: (a) UV–vis absorption and PL spectra in toluene solution (black) compared with the film PL spectrum (blue) and (b) temperature-dependent transient PL in the neat film under a nitrogen atmosphere (inset: prompt component).

The phosphorescence spectrum of TXADO-spiro-DMACF was measured at 77 K (Figure S5). In well agreement with theoretical computation, the T1 and ΔEST are estimated to be 2.86 and 0.28 eV, respectively. Also, we try to record the photo-excited transient decay spectra for the TXADO-spiro-DMACF film. As one can see (Figures b and S6), a prompt fluorescence lifetime (τp) of 3.8 ns is observed in the nanosecond scale. To investigate the potential delayed fluorescence, we then extend the detection time to the millisecond scale under a nitrogen atmosphere. At room temperature, TXADO-spiro-DMACF exhibits a distinct delay component with a delayed fluorescence lifetime (τd) of 106 us, which disappears in the presence of oxygen (Figure S7). As the temperature is down from 300 to 150 K, noticeably, the proportion of the delayed component is found to be gradually decreased, which is a typical feature of TADF emitters.[6,7,10−15]

EL Performance

To evaluate the EL property of TXADO-spiro-DMACF, a nondoped device was fabricated with a configuration of indium tin oxide (ITO)/MoO3 (10 nm)/TAPC (60 nm)/TCTA (5 nm)/TXADO-spiro-DMACF (20 nm)/BmPyPB (30 nm)/LiF (1 nm)/Al (Figure ). Here, MoO3 and LiF are employed as the hole- and electron-injecting materials, respectively; TAPC (4,4′-(cyclohexane-1,1-diyl)bis(N,N-di-p-tolylaniline)) and BmPyPB (3,3″,5,5″-tetra(pyridin-3-yl)-1,1′:3′,1″-terphenyl) act as the hole- and electron-transporting materials, respectively; TCTA is selected as the buffer layer and exciton-blocking layer and TXADO-spiro-DMACF is used independently as the emitting layer.

Figure 4

(a) Nondoped device configuration; (b) molecular structures of used materials.

The EL spectrum, current density–voltage–luminance characteristics, and the luminance dependence of the current efficiency and EQE are plotted in Figure . TXADO-spiro-DMACF displays a bright deep-blue EL with a peak at 444 nm, which is almost independent on the driving voltages, and a long tail between 500 and 600 nm is also observed in the EL spectra, which could be tentatively attributed to the electromer emission.[40] Despite this, the CIE coordinates are (0.16, 0.09), close to the National Television System Committee standard viz. (0.14, 0.08). To our knowledge, this is the first realization of deep-blue (CIE ≤ 0.10) TADF OLEDs on the basis of a nondoped device configuration. In the meantime, a comparable device performance is also attained, giving a maximum luminance of 3120 cd/m2, a current efficiency of 5.3 cd/A, and an EQE of 5.3% (Table ). At a luminance of 200 cd/m2, the EQE further decays to 3.0% (Figure S8). The observed efficiency roll-off could be ascribed to the triplet–triplet annihilation induced by the long delayed lifetime of 106 μs. More interestingly, the deep-blue device can work at an extremely low driving voltage, leading to power efficiency as high as 5.9 lm/W (Figure S9). For instance, the turn-on voltage at 1 cd/m2 is about 2.8 V, even lower than that of the optical band gap of TXADO-spiro-DMACF (3.10 eV). The results suggest the good charge injection and transporting capability and show the great potential of TXADO-spiro-DMACF used for nondoped devices. On the basis of 10 wt % TXADO-spiro-DMACF doped into TCTA as the emitting layer instead of the neat TXADO-spiro-DMACF, we note, the emissive maximum is further blue-shifted from 444 to 428 nm for the doped device together with a peak EQE of 4.3% and CIE coordinates of (0.16, 0.05) (Figure S10 and Table S3).

Figure 5

Nondoped device performance: (a) EL spectra at different driving voltages; (b) current density–voltage–luminance characteristics; and (c) EQE and current efficiency as a function of luminance.

Table 2

Device Performance of TXADO-Spiro-DMACF Compared with the Currently Reported Deep-Blue TADF Emitters with CIE ≤ 0.10

compound	device configuration	Vona [V]	λEL [nm]	CIE (x, y)	EQE [%]	PEb [lm W–1]	CEc [cd A–1]	reference
TXADO-spiro-DMACF	nondoped	2.8	444	(0.16, 0.09)	5.3	5.9	5.3	this work
tDCz-DPS	doped	d	423	(0.15, 0.07)	9.9	d	d	(20)
DTC-mBPSB	doped	d	d	(0.15, 0.08)	5.5	d	4.4	(21)
Cz-TRZ4	doped	3.5	d	(0.15, 0.10)	19.2	d	d	(22)
DABNA-1	doped	d	459	(0.13, 0.09)	13.5	8.3	10.6	(23)
DCzBN3	doped	4.0	428	(0.16, 0.06)	10.3	3.5	5.1	(24)

Turn-on voltage at 1 cd/m2.

Power efficiency.

Current efficiency.

Not reported.

Conclusions

In summary, a spiro-blocking strategy has been newly demonstrated to realize deep-blue and TADF-based nondoped devices. With TXADO-spiro-DMACF as an example, a gentle bathochromic shift associated with a narrowed PL profile, a moderate PLQY and an obvious TADF characteristic are achieved in solid states because of the reduced intermolecular interactions induced by the spiro-blocking. As a result, the TXADO-spiro-DMACF-based nondoped device reveals a deep-blue emission simultaneously with good color purity and high efficiency. This work, we think, will shed light on the development of deep-blue TADF emitters suitable for nondoped OLEDs and subsequently the construction of efficient deep-blue or blue hyperfluorescence systems.

Experimental Section

Measurements and Characterizations

1H NMR and 13C NMR spectra were recorded with a Bruker AVANCE NMR spectrometer. Elemental analysis was performed using a Bio-Rad elemental analysis system. MALDI-TOF mass spectra were performed on an AXIMA CFR MS apparatus (COMPACT). Thermal gravimetric analysis (TGA) and differential scanning calorimetry (DSC) were performed with a PerkinElmer-TGA 7 and PerkinElmer-DSC 7 instrument, respectively, under a nitrogen atmosphere at a heating rate of 10 °C/min. UV–vis absorption and PL spectra were recorded on a PerkinElmer LAMBDA 35 UV–vis spectrometer and PerkinElmer LS 50B spectrofluorometer, respectively. The PLQYs were measured using a quantum yield measurement system (C10027, Hamamatsu Photonics) excited at 360 nm. The transient PL spectra were measured by a HORIBA Jobin Yvon Fluorolog-3 spectrofluorometer. Also, the prompt and delay lifetimes were estimated according to a monoexponential and tri-exponential fittings, respectively. CV curves were recorded on an EG&G 283 Princeton Applied Research potentiostat/galvanostat system. Ferrocene was used as the reference and n-Bu4NClO4 was used as the supporting electrolyte. The HOMO and LUMO energy levels were calculated according to the equation HOMO = −e[Eonset,ox + 4.8] V, LUMO = HOMO + Eg, where Eonset,ox was the onset value of the first oxidation potential, and Eg was the optical band gap estimated from the absorption onset.

Device Fabrication and Testing

All the devices were fabricated on glass substrates precoated with 180 nm ITO with a sheet resistance of 10 Ω/sq. The ITO substrates were degreased in ultrasonic solvent bath and then dried at 120 °C for 30 min. Before loaded into the vacuum deposition chamber, the ITO surface was treated with UV–ozone for 15 min. Then, 10 nm of MoO3 was firstly deposited on the top of ITO substrates, followed by 60 nm TAPC, 5 nm TCTA, 20 nm emitting layer, and 35 nm BmPyPB. Finally, 1 nm of LiF and 120 nm of Al were sequentially deposited to form the cathode. All the layers were grown by thermal evaporation in a high-vacuum system with a pressure less than 5 × 10–4 Pa without breaking vacuum. The organic materials were evaporated at the rate in a range of 1–2 Å/s, and the metal Al was evaporated at the rate of 8–10 Å/s. The overlap between ITO and Al electrodes was 4 mm × 4 mm, which is the active emissive area of the OLED devices. The current–voltage–brightness characteristics were measured by using a set of Keithley source measurement units (Keithley 2400 and Keithley 2000) with a calibrated silicon photodiode. The EL spectra and CIE coordinates were recorded with a PR650 spectra colorimeter. All the measurements were carried out in an ambient atmosphere at room temperature.

Materials

All starting materials for the synthesis were purchased from commercial sources and used as received unless otherwise noted. Tetrahydrofuran (THF) used for the Grignard reaction and toluene used for the Buchwald–Hartwig reaction were dried by sodium and distilled under an argon atmosphere.

Synthetic Procedures

4,4′-Dibromo-2-nitro-1,1′-biphenyl (1)

A mixture of 4,4′-dibromo-1,1′-biphenyl (20 g, 64.10 mmol) and ∼300 mL of acetic acid was heated to 100 °C (colorless transparent solution). Then ∼80 mL of fuming nitric acid was added dropwise. After being stirred for ∼2 h, the solution was cooled to room temperature, poured into ∼2 L of water with a stirring bar. The resulting solid was collected by filtration, dried in vacuum at 70 °C overnight. The crude product was purified by column chromatography to give some bright yellow solid 1 (15.30 g, 67%). 1H NMR (400 MHz, CDCl3): δ 8.03 (d, J = 2.0 Hz, 1H), 7.76 (dd, J = 8.2, 2.0 Hz, 1H), 7.59–7.57 (m, 1H), 7.56–7.54 (m, 1H), 7.29 (d, J = 8.0 Hz, 1H), 7.18–7.16 (m, 1H), 7.15–7.13 (m, 1H).

4,4′-Dibromo-[1,1′-biphenyl]-2-amine (2)

A mixture of 1 (15 g, 42.90 mmol) and ∼400 mL of the mixed solvent (acetic acid/water = 8:1) was heated to 100 °C (colorless transparent solution). A total of ∼12 g of Fe powder was added 4 times in ∼2 h. The solution was cooled to room temperature, poured into ∼2 L of water with a stirring bar. A brown solid was separated by filtration, dried in vacuum at 70 °C overnight to get the crude product 2, which was used in the next step without further purification.

4,4′-Dibromo-2-iodo-1,1′-biphenyl (3)

A mixture of 2 (14 g, 44.50 mmol) and ∼300 mL of hydrochloric acid (6 mol/L) was stirred with a mechanical agitator at 0 °C for 30 min. A solution of NaNO2 (5 g in 100 mL H2O) was added dropwise. After being stirring for 2 h, a solution of KI (48 g, 100 mL H2O) was added dropwise. After 2 h, the temperature was raised to 50 °C for another 2 h. The mixture was poured into ∼2 L water with a stirring bar. A proper amount of Na2SO3 was added into the system to consume the excessive volatile iodine. After filtration and drying, the crude product was further purified by column chromatography to give some colorless (often purple because of the residue I2) solid 3 (12.80 g, 66%). 1H NMR (400 MHz, CDCl3): δ 8.10 (d, J = 1.9 Hz, 1H), 7.58–7.56 (m, 1H), 7.54 (m, 1H), 7.52 (dd, J = 8.2, 1.9 Hz, 1H), 7.20–7.18 (m, 1H), 7.17 (m, 1H), 7.13 (d, J = 8.2 Hz, 1H).

9-(4,4′-Dibromo-[1,1′-biphenyl]-2-yl)-9-hydroxy-9H-thioxanthene 10,10-Dioxide (4)

All the glass containers were heated by a heating gun and cooled to room temperature for 3 times and the reaction was conducted under argon atmosphere. A mixture of ∼1 g LiCl, 1 iodine grain, ∼1 g magnesium pieces were added before ∼100 mL dry THF. A solution of 3 (8.76 g, 20 mmol) and 1 drop of bromoethane (as the initiator) in ∼50 mL of dry THF was added dropwise into the system. After being stirring for 30 min at 60 °C, the solution was transfer into a flask with 9H-thioxanthen-9-one 10,10-dioxide (7.33 g, 30 mmol) through a needle tube connecting the two spaces, stirred at 60 °C for 2 h. The mixture was cooled to room temperature before ∼200 mL of water was added, extracted with 200 mL of dichloromethane twice. The combined organic solution was dried with Na2SO4, filtrated, and removed the solvent in vacuum. The residue was further purified by column chromatography to give some white solid 4 (9.11 g, 50%). 1H NMR (400 MHz, DMSO): δ 8.15 (s, 1H), 7.80 (dd, J = 11.2, 5.0 Hz, 2H), 7.62–7.54 (m, 5H), 7.30 (s, 1H), 7.24 (dd, J = 13.3, 10.8 Hz, 2H), 6.93 (d, J = 8.2 Hz, 2H), 6.83 (d, J = 8.0 Hz, 1H), 5.80 (d, J = 8.1 Hz, 2H).

2,7-Dibromospiro[fluorene-9,9′-thioxanthene] 10′,10′-Dioxide (5)

A mixture of 4 (8 g, 14.11 mmol) and ∼400 mL of acetic acid was heated to 100 °C (colorless transparent solution). ∼50 mL of hydrochloric acid (6 mol/L) was added slowly. After being stirring overnight, the system was cooled to room temperature resulting in some slightly yellow precipitation. The solid was collected by filtration and washed with water three times, and dried in vacuum at 70 °C overnight. The crude product was further purified by column chromatography to give some white solid 5 (4.77 g, 62%). 1H NMR (400 MHz, CDCl3): δ 8.26 (dd, J = 8.0, 0.9 Hz, 2H), 7.68 (d, J = 8.2 Hz, 2H), 7.59–7.49 (m, 4H), 7.46 (d, J = 1.5 Hz, 2H), 7.35–7.29 (m, 2H), 6.56 (d, J = 7.8 Hz, 2H).

TXADO-Spiro-DMACF (6)

The reaction was conducted under an argon atmosphere. A mixture of 5 (1.20 g, 2.20 mmol), 9,9-dimethyl-9,10-dihydroacridine (1.10 g, 5.30 mmol), Pd2(dba)3 (0.050 g, 0.053 mmol), P(t-Bu)3HBF4 (0.064 g, 0.22 mmol), and t-BuONa (2.10 g, 22 mmol) was added before ∼50 mL trimethylbenzene and heated to 135 °C overnight. The system was cooled to room temperature and most of the solvent was removed by reduced pressure distillation. The residue was further purified by column chromatography to give some white solid 6 (1.76 g, 96%). Then, the product was recrystallized in a mixed solvent of dichloromethane and n-hexane twice. 1H NMR (400 MHz, CDCl3): δ 8.18–8.13 (m, 4H), 7.50–7.43 (m, 4H), 7.42–7.33 (m, 8H), 6.92–6.83 (m, 10H), 6.19–6.11 (m, 4H), 1.61 (s, 12H). 13C NMR (101 MHz, CDCl3): δ 156.32, 142.17, 140.79, 139.61, 138.65, 136.92, 132.72, 132.37, 130.28, 129.25, 128.87, 128.68, 126.20, 125.24, 123.80, 122.79, 120.78, 113.73, 35.98, 30.98. MALDI-TOF MS: 794.3 [M+]. Anal. Calcd for C55H42N2O2S: C, 83.09; H, 5.33; N, 3.52. Found: C, 82.94; H, 5.44; N, 3.43.