Deep-Blue Triplet–Triplet Annihilation Organic Light-Emitting Diode (CIEy ≈ 0.05) Using Tetraphenylimidazole and Benzonitrile Functionalized Anthracene/Chrysene Emitters

Herein, new deep-blue triplet-triplet annihilation (TTA) molecules, namely 4-(10-(4-(1,4,5-triphenyl-1H-imidazol-2-yl)phenyl)anthracen-9-yl)benzonitrile (TPIAnCN) and 4-(12-(4-(1,4,5-triphenyl-1H-imidazol-2-yl)phenyl)chrysen-6-yl)benzonitrile (TPIChCN), are designed, synthesized, and investigated as emitters for organic light-emitting diodes (OLED). TPIAnCN and TPIChCN are composed of polyaromatic hydrocarbons of anthracene (An) and chrysene (Ch) as the cores functionalized with tetraphenylimidazole (TPI) and benzonitrile (CN) moieties, respectively. The experimental and theoretical results verify their excellent thermal properties, photophysical properties, as well as electrochemical properties. Particularly, their emissions are in the deep blue region, with TTA emissions being observed in their thin films. By utilization of these molecules as emitters, deep blue TTA OLEDs with CIE coordinates of (0.15, 0.05), high external quantum efficiency of 6.84%, and high exciton utilization efficiency (ηs) of 48% were fabricated. This result manifests the potential use of chrysene as an alternate building block to formulate new TTA molecules for accomplishing high-performance TTA OLEDs.


Introduction
Organic light-emitting diodes (OLEDs) have been widely studied since the publication of the pioneering report in 1978 by Tang et al. [1] owing to their attractive characteristics (transparent lighting panels, high brightness, color tuneability, as well as flexibility) and potential applications in the new generation of display and solid-state lighting technologies [2][3][4][5][6]. In particular, blue light-emitting OLEDs are indispensable as one of the three primary colors (blue, green, and red) for full-color display. However, in terms of device performance, blue OLEDs still lag behind both red and green OLEDs to date. This is because of the extrinsically wide energy band gap of blue emitters, complicating the molecular design and the discovery of the well-matched hole transporting layer or electron transporting layer [7][8][9]. Hence, the development of blue emitters for blue OLEDs is very fascinating and challenging to study for the improvement of device performance. Due to the limit of efficiency for blue fluorescent OLEDs with the theoretical maximum external quantum efficiency (EQE max ) of 5%, harvesting only singlet excitons [10][11][12][13][14], there are currently several blue OLED mechanisms for enhancing the device performance to harvest the triplet excitons, such as phosphorescence, thermally activated delayed fluorescence (TADF) [15][16][17][18], hybridized local charge transfer (HLCT) [19,20], and triplet-triplet annihilation (TTA) [21,22]. These mechanisms have achieved EQEs of higher than 5%. However, the metal-containing blue phosphorescent materials barely give deep-blue emissions attributable to their radiative metal-ligand charge transfer (MLCT) nature. Moreover, these
For a structural conformation and a deeper understanding of photophysical properties in the solid state of TPIAnCN and TPIChCN, an X-ray crystallographic analysis was performed. Their crystals for single-crystal X-ray diffraction (SC-XRD) studies were crystallized by using solvent and anti-solvent evaporation methods in a mixture of CHCl3/MeOH at room temperature. The structural refinement of TPIAnCN presented in the monoclinic crystal system with the space group of P21/c (a = 6.0526(4) Å , b = 14.6484 (9) Å, c = 38.294(3) Å with β = 93.888 (3) Å. Whereas TPIChCN presented in the triclinic crystal system with the space group of P-1 (a = 10.0312(6) Å , b = 10.7769(7) Å , c = 18.8593(11) Å with β = 101.607 (2) Å and γ 107.107 (2) Å. The complete crystallographic information and molecular packing are given in Table S1 and Figures S2 and S3. As shown in Figure 1, both crystal structures showed their intermolecular interaction, mainly including CH-π and ππ interactions, fashioned face-to-face packing between aromatic planar units of anthracene for TPIAnCN and chrysene for TPIChCN. The interactions affected the fluorescent quenching phenomena in both solid powder and neat film by the continuous π-π interaction of anthracene and chrysene-cored structure (Figure 1), which caused a decrease in the PL quantum yield of the J-aggregation. Furthermore, a lone pair-π interaction (lp-π) was also found between the CN group and phenyl ring on the imidazole unit of TPIChCN, which caused the additional fluorescent quenching by electronic charge transfer [54]. To tackle both π-stacking and lp-π interaction which could hinder the OLED performance, a doped-layered OLED structure is introduced and studied.
For a structural conformation and a deeper understanding of photophysical properties in the solid state of TPIAnCN and TPIChCN, an X-ray crystallographic analysis was performed. Their crystals for single-crystal X-ray diffraction (SC-XRD) studies were crystallized by using solvent and anti-solvent evaporation methods in a mixture of CHCl 3 /MeOH at room temperature. The structural refinement of TPIAnCN presented in the monoclinic crystal system with the space group of P21/c (a = 6.0526(4) Å, b = 14.6484(9) Å, c = 38.294(3) Å with β = 93.888 (3) Å). Whereas TPIChCN presented in the triclinic crystal system with the space group of P-1 (a = 10.0312(6) Å, b = 10.7769(7) Å, c = 18.8593(11) Å with β = 101.607 (2) Å and γ 107.107 (2) Å). The complete crystallographic information and molecular packing are given in Table S1 and Figures S2 and S3. As shown in Figure 1, both crystal structures showed their intermolecular interaction, mainly including CH-π and π-π interactions, fashioned face-to-face packing between aromatic planar units of anthracene for TPIAnCN and chrysene for TPIChCN. The interactions affected the fluorescent quenching phenomena in both solid powder and neat film by the continuous π-π interaction of anthracene and chrysene-cored structure (Figure 1), which caused a decrease in the PL quantum yield of the J-aggregation. Furthermore, a lone pair-π interaction (lp-π) was also found between the CN group and phenyl ring on the imidazole unit of TPIChCN, which caused the additional fluorescent quenching by electronic charge transfer [54]. To tackle both π-stacking and lp-π interaction which could hinder the OLED performance, a doped-layered OLED structure is introduced and studied. Single crystal structures (thermal ellipsoids at the 50% probability) and crystal packing of (a) TPIAnCN and (b) TPIChCN (C-H-π (green), π-π (orange), and lp-π (pink) interactions).
For a deeper insight into the structural and electronic properties of TPIAnCN and TPIChCN, the theoretical calculations were performed using the B3LYP/6-31G(d,p) density functional theory (DFT) in CH2Cl2. As depicted in Figure 2a, their optimized structures showed a twisting conformation with the angles between anthracene and adjacent phenyl rings (71° and 57°) being larger than those of chrysene (50° and 56°). Hence, such highly twisted geometry would disrupt the molecular packing in the solid state to some degree, confining the fluorescence emission in the deep blue region. In frontier molecular orbitals, electrons in the highest occupied molecular orbital (HOMO) of TPIAnCN localized largely on the anthracene ring with a partial distribution on the TPI unit, whereas in the HOMO of TPIChAN, localization of electrons was observed mainly on the TPI moiety. On the other hand, the excited electrons in the lowest unoccupied molecular orbital (LUMO) of both molecules was delocalized over the conjugated backbones of 4-cyanophenyl anthracene for TPIAnCN and 4-cyanophenyl chrysene for TPIChAN. Additionally, the energy levels of both excited states (S and T) were measured using the time-dependent (TD)-DFT calculations using the B3LYP/6-31G(d,p) method. As illustrated in Figure 2b, the S1 and T1 excited energy levels of TPIAnCN and TPIChCN are calculated to be 2.42 and 2.53 eV, 1.67 eV, and 1.45 eV, respectively. As a result of their estimated ΔEST values of 0.75-1.45 eV, the difficulty is assured for both molecules to express the T1 to S1 reverse intersystem crossing (RISC) via the TADF mechanism. However, the two-triplet fusion energy levels (2T1) of the two molecules agree well with the principle of 2T1 > S1, verifying that the up-conversion of T1 into S1 via a TTA mechanism is achievable in both molecules [55,56]. Single crystal structures (thermal ellipsoids at the 50% probability) and crystal packing of (a) TPIAnCN and (b) TPIChCN (C-H-π (green), π-π (orange), and lp-π (pink) interactions).
For a deeper insight into the structural and electronic properties of TPIAnCN and TPIChCN, the theoretical calculations were performed using the B3LYP/6-31G(d,p) density functional theory (DFT) in CH 2 Cl 2 . As depicted in Figure 2a, their optimized structures showed a twisting conformation with the angles between anthracene and adjacent phenyl rings (71 • and 57 • ) being larger than those of chrysene (50 • and 56 • ). Hence, such highly twisted geometry would disrupt the molecular packing in the solid state to some degree, confining the fluorescence emission in the deep blue region. In frontier molecular orbitals, electrons in the highest occupied molecular orbital (HOMO) of TPIAnCN localized largely on the anthracene ring with a partial distribution on the TPI unit, whereas in the HOMO of TPIChAN, localization of electrons was observed mainly on the TPI moiety. On the other hand, the excited electrons in the lowest unoccupied molecular orbital (LUMO) of both molecules was delocalized over the conjugated backbones of 4-cyanophenyl anthracene for TPIAnCN and 4-cyanophenyl chrysene for TPIChAN. Additionally, the energy levels of both excited states (S and T) were measured using the time-dependent (TD)-DFT calculations using the B3LYP/6-31G(d,p) method. As illustrated in Figure 2b, the S 1 and T 1 excited energy levels of TPIAnCN and TPIChCN are calculated to be 2.42 and 2.53 eV, 1.67 eV, and 1.45 eV, respectively. As a result of their estimated ∆E ST values of 0.75-1.45 eV, the difficulty is assured for both molecules to express the T 1 to S 1 reverse intersystem crossing (RISC) via the TADF mechanism. However, the two-triplet fusion energy levels (2T 1 ) of the two molecules agree well with the principle of 2T 1 > S 1 , verifying that the up-conversion of T 1 into S 1 via a TTA mechanism is achievable in both molecules [55,56].   Table 1. The UV-Vis absorption spectra in toluene of TPIAnCN and TPIChCN unveiled two obvious absorption bands around 260-270 nm originating from the aromatic rings [57] and 350-405 nm assigned to the π-π* transition of anthracene moiety and chrysene moiety of TPIAnCN and TPIChCN, respectively ( Figure 3a) [50,58]. The optical bandgaps (Eg opt ) of TPIAnCN and TPIChCN were estimated from their absorption onsets to be 2.85 and 3.06 eV, respectively. In the solution, both molecules exhibited a strong fluorescence emission in a deep blue region. The photoluminescence (PL) spectra of TPIAnCN and TPIChCN were located at 441 and 432 nm, respectively, whereas their PL emission peaks in neat films showed a slight redshift to 463 and 462 nm for TPIAnCN and TPIChCN, owing to the enhancement of intermolecular interactions in the solid film, respectively ( Figure 3b). The absolute PL quantum yields (ΦPL) measured by an integrating sphere of TPIAnCN and TPIChCN in toluene were evaluated to be 60% and 59%, respectively. On the one hand, in the neat film, their ΦPL values of TPIAnCN and TPIChCN were dropped to 21% and 49%, respectively, signifying that intermolecular π-π interaction exists in the film state causing a fluorescence quenching. The high ΦPL values of the molecules in film states could be restored by doping in 4′-bis(N-carbazolyl)-1,1′-biphenyl (CBP) film. With the optimum doping concentration of 5 wt%, the ΦPL values of TPIAnCN and TPIChCN in film state were measured to be 71% and 73%, respectively. As shown in Figure 3b, the PL spectra of the doped films have shown slight blue-shift compared to the neat films with no emission peaks of the CBP host matrix being spotted, indicating a complete energy transfer from the CBP host to the dopants. A further study considering the PL lifetime measurements revealed that both compounds in neat film displayed mono-exponential decay profiles with the range of 0.96-1.41 ns (Figure 3c and Table 1), suggesting their PL emissions stem from the singlet excited state.    Figure S4) and the key data are listed in Table 1. The UV-Vis absorption spectra in toluene of TPIAnCN and TPIChCN unveiled two obvious absorption bands around 260-270 nm originating from the aromatic rings [57] and 350-405 nm assigned to the π-π* transition of anthracene moiety and chrysene moiety of TPIAnCN and TPIChCN, respectively ( Figure 3a) [50,58]. The optical bandgaps (E g opt ) of TPIAnCN and TPIChCN were estimated from their absorption onsets to be 2.85 and 3.06 eV, respectively. In the solution, both molecules exhibited a strong fluorescence emission in a deep blue region. The photoluminescence (PL) spectra of TPIAnCN and TPIChCN were located at 441 and 432 nm, respectively, whereas their PL emission peaks in neat films showed a slight redshift to 463 and 462 nm for TPIAnCN and TPIChCN, owing to the enhancement of intermolecular interactions in the solid film, respectively ( Figure 3b). The absolute PL quantum yields (Φ PL ) measured by an integrating sphere of TPIAnCN and TPIChCN in toluene were evaluated to be 60% and 59%, respectively. On the one hand, in the neat film, their Φ PL values of TPIAnCN and TPIChCN were dropped to 21% and 49%, respectively, signifying that intermolecular π-π interaction exists in the film state causing a fluorescence quenching. The high Φ PL values of the molecules in film states could be restored by doping in 4 -bis(N-carbazolyl)-1,1 -biphenyl (CBP) film. With the optimum doping concentration of 5 wt%, the Φ PL values of TPIAnCN and TPIChCN in film state were measured to be 71% and 73%, respectively. As shown in Figure 3b, the PL spectra of the doped films have shown slight blue-shift compared to the neat films with no emission peaks of the CBP host matrix being spotted, indicating a complete energy transfer from the CBP host to the dopants. A further study considering the PL lifetime measurements revealed that both compounds in neat film displayed mono-exponential decay profiles with the range of 0.96-1.41 ns (Figure 3c and Table 1), suggesting their PL emissions stem from the singlet excited state. films. e Transient PL decay time. f Determined by TGA and DSC at a heating rate of 10 °C min −1 under N2 flow. g Obtained from CV measurement at a scan rate of 50 mV s −1 . h Calculated from the onset of absorption spectra of thin films: Eg opt = 1240/λonset. i HOMO obtained from AC-2 of the neat film and LUMO (eV) = HOMO + Eg opt . Additionally, TTA-induced delay fluorescence of TPIAnCN and TPIChCN were also investigated. The triplet energies of TPIAnCN and TPIChCN were measured using a room temperature triplet state spectroscopic measurement technique [59,60]. TPIAnCN and TPIChCN 2 wt% doped in poly(4-bromostyrene) (PBS) films covered by EX-CEVALTM film were analyzed by time-resolved emission spectroscopy (TRES) ( Figure  4). As depicted in Figure 4a The PL bands at the low wavelength region matched well with their corresponding prompt PL emissions, verifying delayed PL from the S1 state. The PL bands at the longer wavelengths were assignable to their phosphorescence (Ph) emissions. Hence, the S1 state and T1 state energies were calculated from the onsets of those delayed PL and Ph spectra to be 3.05 and 1.80 eV for TPIAnCN, and 3.39 and 2.38 eV for TPIChCN, respectively [50,61]. TTA is considered to be the most plausible mechanism for the observed delayed PL among a few possible mechanisms. Both molecules could produce additional singlet excitons through the triplet fusion process since the energy levels fulfilled the conditions of 2T1 > S1 for TTA-based molecules. Furthermore, the energy gaps between S1 and T1 of greater than 1.25 eV would substantially limit the T1-to-S1 reverse intersystem crossing (RISC) process by way of the TADF mechanism [62].  Additionally, TTA-induced delay fluorescence of TPIAnCN and TPIChCN were also investigated. The triplet energies of TPIAnCN and TPIChCN were measured using a room temperature triplet state spectroscopic measurement technique [59,60]. TPIAnCN and TPIChCN 2 wt% doped in poly(4-bromostyrene) (PBS) films covered by EXCEVALTM film were analyzed by time-resolved emission spectroscopy (TRES) ( Figure 4). As depicted in Figure 4a,b, the TRES maps are composed of two-component emission maps in the ranges of 400-550 nm and 700-750 nm for TPIAnCN and 350-470 nm and 530-670 nm for TPIChCN. The integrated TRES slices of TPIAnCN (PL@1.3 ms) and TPIChCN (PL@4 ms) displayed two PL emission bands (Figure 4c,d). The PL bands at the low wavelength region matched well with their corresponding prompt PL emissions, verifying delayed PL from the S 1 state. The PL bands at the longer wavelengths were assignable to their phosphorescence (Ph) emissions. Hence, the S 1 state and T 1 state energies were calculated from the onsets of those delayed PL and Ph spectra to be 3.05 and 1.80 eV for TPIAnCN, and 3.39 and 2.38 eV for TPIChCN, respectively [50,61]. TTA is considered to be the most plausible mechanism for the observed delayed PL among a few possible mechanisms. Both molecules could produce additional singlet excitons through the triplet fusion process since the energy levels fulfilled the conditions of 2T 1 > S 1 for TTA-based molecules. Furthermore, the energy gaps between S 1 and T 1 of greater than 1.25 eV would substantially limit the T 1 -to-S 1 reverse intersystem crossing (RISC) process by way of the TADF mechanism [62]. The thermal properties of TPIAnCN and TPIChCN were examined by thermal gravimetric analysis (TGA) and differential scanning calorimetry (DSC) under a nitrogen flow. As shown in Figure 5a, both compounds exhibit high decomposition temperature (Td) TGA traces with the Td at 5% weight loss over 500 °C (Table 1), suggesting high thermal stability. The 2nd heating DSC thermograms of TPIAnCN and TPIChCN display only sharp endothermic peaks corresponding to melting temperatures (Tm) of 350, and 364 °C, respectively. No glass transition temperature was observed in both cases, implying that they are crystalline solids. The outstanding thermal stability of TPIAnCN and TPIChCN is critical to achieving a high device performance given that they can withstand high temperatures and will not easily decompose in the device fabrication process using thermal evaporation techniques. The electrochemical properties of TPIAnCN and TPIChCN were investigated using cyclic voltammetry (CV) at a scan rate of 50 mV s −1 in dry CH2Cl2 solution using 0.05 M of n-Bu4NPF6 as a supporting electrolyte. As described in Figure 5b, CV traces show two quasi-reversible oxidation waves ( Table 1). The first oxidation process of both compounds appeared at the same half-wave potential (E1/2) of 1.32 V which could be associated with the oxidation of the TPI unit as observed in the DFT calculation results. The HOMO energy levels of TPIAnCN and TPIChCN in the thin film were determined by photoelectron yield spectroscopy (AC-2) in air to be −5.77 and −5.79 eV, respectively ( Figure S5). The lowest unoccupied molecular orbital (LUMO) energy levels were calculated from the HOMO values and the optical band gaps (Eg opt ) by using the equation LUMO   The thermal properties of TPIAnCN and TPIChCN were examined by thermal gravimetric analysis (TGA) and differential scanning calorimetry (DSC) under a nitrogen flow. As shown in Figure 5a, both compounds exhibit high decomposition temperature (T d ) TGA traces with the T d at 5% weight loss over 500 • C (Table 1), suggesting high thermal stability. The 2nd heating DSC thermograms of TPIAnCN and TPIChCN display only sharp endothermic peaks corresponding to melting temperatures (T m ) of 350, and 364 • C, respectively. No glass transition temperature was observed in both cases, implying that they are crystalline solids. The outstanding thermal stability of TPIAnCN and TPIChCN is critical to achieving a high device performance given that they can withstand high temperatures and will not easily decompose in the device fabrication process using thermal evaporation techniques. The electrochemical properties of TPIAnCN and TPIChCN were investigated using cyclic voltammetry (CV) at a scan rate of 50 mV s −1 in dry CH 2 Cl 2 solution using 0.05 M of n-Bu 4 NPF 6 as a supporting electrolyte. As described in Figure 5b, CV traces show two quasi-reversible oxidation waves ( Table 1). The first oxidation process of both compounds appeared at the same half-wave potential (E 1/2 ) of 1.32 V which could be associated with the oxidation of the TPI unit as observed in the DFT calculation results. The HOMO energy levels of TPIAnCN and TPIChCN in the thin film were determined by photoelectron yield spectroscopy (AC-2) in air to be −5.77 and −5.79 eV, respectively ( Figure S5). The lowest unoccupied molecular orbital (LUMO) energy levels were calculated from the HOMO values and the optical band gaps (E g opt ) by using the equation LUMO (eV) = HOMO + E g opt . The LUMO energy levels for TPIAnCN and TPIChCN were −2.92 and −2.73 eV, respectively.  Figure 6. To achieve the optimal device EL performance, MoO3 as hole injection layer, TAPC as holetransporting layer, TCTA as exciton blocking layer, TmPyPB as electron transporting layer and hole blocking layer, and LiF as electron injection layer were employed. TPIAnCN and TPIChCN 5wt% doped in CBP were used as EML. A combination of suitable hole mobility and HOMO level of TAPC (−5.50 eV) and suitable electron mobility and LUMO level of TmPyPB (−2.75 eV) will conceivably confine the exciton recombination zone width in EML. The characteristic curves and device data are shown in Figure 7 and Table 2. The OLEDs based on TPIAnCN and TPIChCN demonstrated intense blue emission color with the EL emission peaks at 438 and 431 nm with CIE coordinates of (0.15, 0.07) and (0.15, 0.05), respectively. Obviously, both devices delivered a deep blue light close to the HDTV standard blue color. The EL spectra of the two devices matched well with the PL spectra of the corresponding emitters in the thin films. Remarkably, the EL emissions were rather stable with no emission peaks of TAPC (374 nm), TCTA (390 nm) [63], CBP (390 nm) [64], or TmPyPB (471 nm) [64] being seen under the whole range of applied voltages (6-10 V) ( Figure S6). This suggests an efficient charge injection and recombination in the EML, and excimer emission, as well as exciplex emissions at the interfaces of EML/TmPyPB and TCTA/EML, are effectively controlled. Both devices demonstrated low turn-on voltage (Von) of 3.2 and 3.4 V for TPIAnCN and TPIChCN, respectively, indicating effective charge injection and recombination in EML. Among the two, TPIAnCN-based OLED exhibited the highest EL performance with a maximum external quantum efficiency (EQEmax) of 6.84%, whereas TPIChCN-based OLED showed a slightly lower EQEmax of 4.28%. However, in terms of color purity, the device based on TPIChCN with CIE coordinates of (0.15, 0.05) was in deeper blue emission color than TPIAnCN-based OLED. This suggests that the chrysene core is a deeper blue-emitting unit than the anthracene core.   Figure 6. To achieve the optimal device EL performance, MoO 3 as hole injection layer, TAPC as holetransporting layer, TCTA as exciton blocking layer, TmPyPB as electron transporting layer and hole blocking layer, and LiF as electron injection layer were employed. TPIAnCN and TPIChCN 5 wt% doped in CBP were used as EML. A combination of suitable hole mobility and HOMO level of TAPC (−5.50 eV) and suitable electron mobility and LUMO level of TmPyPB (−2.75 eV) will conceivably confine the exciton recombination zone width in EML. The characteristic curves and device data are shown in Figure 7 and Table 2. The OLEDs based on TPIAnCN and TPIChCN demonstrated intense blue emission color with the EL emission peaks at 438 and 431 nm with CIE coordinates of (0.15, 0.07) and (0.15, 0.05), respectively. Obviously, both devices delivered a deep blue light close to the HDTV standard blue color. The EL spectra of the two devices matched well with the PL spectra of the corresponding emitters in the thin films. Remarkably, the EL emissions were rather stable with no emission peaks of TAPC (374 nm), TCTA (390 nm) [63], CBP (390 nm) [64], or TmPyPB (471 nm) [64] being seen under the whole range of applied voltages (6-10 V) ( Figure S6). This suggests an efficient charge injection and recombination in the EML, and excimer emission, as well as exciplex emissions at the interfaces of EML/TmPyPB and TCTA/EML, are effectively controlled. Both devices demonstrated low turn-on voltage (V on ) of 3.2 and 3.4 V for TPIAnCN and TPIChCN, respectively, indicating effective charge injection and recombination in EML. Among the two, TPIAnCN-based OLED exhibited the highest EL performance with a maximum external quantum efficiency (EQE max ) of 6.84%, whereas TPIChCN-based OLED showed a slightly lower EQE max of 4.28%. However, in terms of color purity, the device based on TPIChCN with CIE coordinates of (0.15, 0.05) was in deeper blue emission color than TPIAnCN-based OLED. This suggests that the chrysene core is a deeper blue-emitting unit than the anthracene core.  Figure 6. To achieve the optimal device EL performance, MoO3 as hole injection layer, TAPC as holetransporting layer, TCTA as exciton blocking layer, TmPyPB as electron transporting layer and hole blocking layer, and LiF as electron injection layer were employed. TPIAnCN and TPIChCN 5wt% doped in CBP were used as EML. A combination of suitable hole mobility and HOMO level of TAPC (−5.50 eV) and suitable electron mobility and LUMO level of TmPyPB (−2.75 eV) will conceivably confine the exciton recombination zone width in EML. The characteristic curves and device data are shown in Figure 7 and Table 2. The OLEDs based on TPIAnCN and TPIChCN demonstrated intense blue emission color with the EL emission peaks at 438 and 431 nm with CIE coordinates of (0.15, 0.07) and (0.15, 0.05), respectively. Obviously, both devices delivered a deep blue light close to the HDTV standard blue color. The EL spectra of the two devices matched well with the PL spectra of the corresponding emitters in the thin films. Remarkably, the EL emissions were rather stable with no emission peaks of TAPC (374 nm), TCTA (390 nm) [63], CBP (390 nm) [64], or TmPyPB (471 nm) [64] being seen under the whole range of applied voltages (6-10 V) ( Figure S6). This suggests an efficient charge injection and recombination in the EML, and excimer emission, as well as exciplex emissions at the interfaces of EML/TmPyPB and TCTA/EML, are effectively controlled. Both devices demonstrated low turn-on voltage (Von) of 3.2 and 3.4 V for TPIAnCN and TPIChCN, respectively, indicating effective charge injection and recombination in EML. Among the two, TPIAnCN-based OLED exhibited the highest EL performance with a maximum external quantum efficiency (EQEmax) of 6.84%, whereas TPIChCN-based OLED showed a slightly lower EQEmax of 4.28%. However, in terms of color purity, the device based on TPIChCN with CIE coordinates of (0.15, 0.05) was in deeper blue emission color than TPIAnCN-based OLED. This suggests that the chrysene core is a deeper blue-emitting unit than the anthracene core.    To reveal the EL mechanism in these devices, transient EL measurements were performed. As shown in Figure 8, two components of prompt EL decay from rapid emission of S1 and delayed EL from the TTA channel are covered in the transient EL profiles of the two OLEDs. The delayed component ratio slightly decreased when the driving voltages increased indicating the presence of the TTA emission. The TTA process is extremely effective at a low driving voltage since this can create rich delayed components, giving rise to the improvement in the device performance. However, as the driving voltages increased, the delayed components decreased because the triplet excitons are quenched, resulting in decreased EQE [48].  To reveal the EL mechanism in these devices, transient EL measurements were performed. As shown in Figure 8, two components of prompt EL decay from rapid emission of S 1 and delayed EL from the TTA channel are covered in the transient EL profiles of the two OLEDs. The delayed component ratio slightly decreased when the driving voltages increased indicating the presence of the TTA emission. The TTA process is extremely effective at a low driving voltage since this can create rich delayed components, giving rise to the improvement in the device performance. However, as the driving voltages increased, the delayed components decreased because the triplet excitons are quenched, resulting in decreased EQE [48]. of S1 and delayed EL from the TTA channel are covered in the transient EL profiles of the two OLEDs. The delayed component ratio slightly decreased when the driving voltages increased indicating the presence of the TTA emission. The TTA process is extremely effective at a low driving voltage since this can create rich delayed components, giving rise to the improvement in the device performance. However, as the driving voltages increased, the delayed components decreased because the triplet excitons are quenched, resulting in decreased EQE [48]. To further realize the emitter's behavior in the device, hole mobilities of the EMLs were estimated using the space-charge-limited current (SCLC) measurements of the holeonly devices (HOD) [65]. The hole-only devices were fabricated with the structure of ITO/MoO 3 (10 nm)/EML (TPIAnCN and TPIChCN 5 wt% in CBP) (100 nm)/MoO 3 (10 nm)/Al (100 nm). The current density-voltage (J-V) plots of HODs are shown in Figure 7d. Accordingly, TPIAnCN and TPIChCN emitters possess a relatively high and fast current with hole mobilities of 2.02 × 10 −4 and 1.84 × 10 −4 cm 2 V −1 s −1 , respectively. Such high hole mobility could contribute to widening the recombination zone in the EML, resulting in a longer device lifetime as well as lower driving voltages. Consequently, the superior EL performance of the TPIAnCN-based OLED could be ascribed to a combination of a high thin film Φ PL , high hole mobility, and suitable HOMO/LUMO levels of the TPIAnCN emitter.
Furthermore, the singlet exciton utilization efficiency (η s ) was calculated following the equation of EQE = η out × η rec × η s × Φ PL , where the light outcoupling efficiency (η out ) is 0.2 for glass substrate, and charge recombination efficiency (η rec ) is estimated to be 1 [66]. As a result of the Φ PL values of 71% for TPIAnCN emitter and 73% for TPIChCN emitter, the corresponding η s values of TPIAnCN and TPIChCN-based devices were estimated to be 48% and 36%, respectively. These η s surpass the statistical limit of 25% of traditional fluorescence emitters, verifying that both TPIAnCN and TPIChCN are TTA emitters. Therefore, the superb EL performance of both OLEDs could be ascribed to a combination of a high fluorescence feature and good hole-transporting property of the emitters as well as the TTA process in the device, in which all these properties are instigated by π-interactions of the polyaromatic rings (anthracene and chrysene) in the molecule in solid-state.

Conclusions
In summary, new triplet-triplet annihilation (TTA)-based deep blue emitters (TPIAnCN and TPIChCN) were designed and characterized. They demonstrated deep blue emission with peaks of around 430-440 nm with high fluorescence. TPIAnCN and TPIChCN were successfully fabricated as emitters in OLEDs. All devices displayed deep blue electroluminescence (EL) spectra with good color purity and CIE coordinates in high-definition television (HDTV) regions. In particular, TPIAnCN-based OLED reached the maximum EQE of 6.84% with CIE coordinates of (0.15, 0.07). The TPIChCN-based device represented one of the deepest blue-emitting TTA-OLEDs with CIE coordinates of (0.15, 0.05). This work not only comprehensively demonstrates the successful use of chrysene as an alternative building block to develop new TTA molecules for achieving high-performance deep blue TTA OLEDs, but also provides a novel design strategy and could potentially be beneficial for exploring high-performance blue OLEDs in the future.

Materials and Methods
All the reagents and solvents obtained from suppliers were used without further purification. The 1 H-and 13 C-NMR spectra were recorded using Bruker (Billerica, MA, USA) AVANCE III HD 600 MHz spectrometer with CDCl 3 as a solvent. The high-resolution mass spectra were analyzed using APCI-TOF Bruker (Billerica, MA, USA) Compact mass spectrometer or Bruker (Billerica, MA, USA) Autoflex Speed TM mass spectrometer. UV-Vis absorption spectra both in solution and thin film were measured using PerkinElmer (Waltham, MA, USA) model Lambda 1050 spectrophotometer. Photoluminescence spectra, lifetime, and TRES measurements both in solution and thin film were analyzed with an Edinburgh (Livingston, UK) FLS980 spectrophotometer. Absolute PL quantum yield was measured by the integrating sphere. Photoelectron spectroscopy (AC-2) was measured by Riken-Keiki (Itabashi, Tokyo, Japan) ultraviolet photoelectron spectrometer AC-2 in air. Thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC) analysis were performed using a Rigaku (Akishima, Tokyo, Japan), model Thermoplus EV02 TG-DTA8122 and PerkinElmer ((Waltham, MA, USA), model DSC8500 with a heating rate of 10 • C min −1 under N 2 gas flow. Cyclic voltammetry (CV) analyses were carried out using Metrohm (Ionenstrasse, Herisau, Switzerland) Autolab potentiostat PGSTAT 101 in CH 2 Cl 2 under Ar atmosphere at a scan rate of 50 mV s −1 (platinum as a counter electrode, glassy carbon as a working electrode, Ag/AgCl as a reference electrode, n-Bu 4 NPF 6 as a supporting electrolyte. Melting points were measured using a Krüss (Borsteler Chaussee, Hamburg, Germany) KSP1N melting point meter and were uncorrected. Single crystal X-ray diffraction (SC XRD) was collected using a Bruker (Billerica, MA, USA) D8 Venture spectrometer at 190 K (Mo K α = 0.7107 Angstrom). The crystal refinement was calculated using APEX4, PLATON (100117), and OLEX2 software. All quantum chemical calculations were based on density functional theory (DFT) and performed with Gaussian 16 program package [67]. The ground state geometries, HOMO and LUMO distributions, and HOMO and LUMO energy levels were calculated by the B3LYP/6-31G(d,p) level of theory. The energy level of the singlet (S) excited and triplet (T) excited states were computed using time-dependent (TD)-DFT calculation with the B3LYP/6-31G(d,p) method.