Ultrastable green top-emitting OLEDs approaching BT.2020 enabled by narrowband phosphorescent emitters Authors: Feng Zhan, Yunyang Yuan, Kewei Xu, Kongwu Wu, Yuanbin She, Guijie Li Published: 2026 ABSTRACT Highly stable and efficient organic light-emitting diodes (OLEDs) with high color purity are critical for ultrahigh-definition display applications. However, achieving these characteristics is hindered by the emitter bandwidth and device stability. Herein, we present the design of tetradentate green Pt(II) emitters with a rigid three-dimensional structure via extended π-conjugation and bulky substituents. This molecular engineering approach effectively suppresses excited-state geometric distortion, resulting in a small full width at half-maximum (FWHM) of 16.4 nanometers in toluene. When the emitters were used in a microcavity-enhanced top-emitting device, the FWHM was further reduced to 14.8 nanometers by controlling constructive and destructive interferences. The top-emitting device showed a Broadcast Service Television 2020 (BT.2020) green color [Commission Internationale de l’Éclairage (CIE): 0.22, 0.75] and a high current efficiency of up to 174 candelas per ampere. Most notably, the device exhibited a record-breaking operational stability, exceeding that of the commercial LED/OLED/MicroLED (micro–light-emitting diode) standard, achieving an unprecedented lifetime to 90% of the initial luminance of 540,100 hours at 1000 candelas per square meter. INTRODUCTION In the contemporary digital landscape, the development of wide-color-gamut displays is of paramount importance for delivering vivid, accurate, and lifelike visual images across various applications. To achieve this goal, the International Telecommunication Union established the Broadcast Service Television 2020 (BT.2020), defining a broader set of primary colors (1). Meeting the BT.2020 standard requires emission peaks within specific wavelength regions and sufficiently small full width at half-maximum (FWHM) values. Quantum dot light-emitting diodes (QD-LEDs) exhibit extremely small FWHM values (∼20 to 30 nm), owing to their nanoscale size control (2, 3); however, the high production costs and storage challenges of these devices hinder their large-scale application (4). Conversely, organic light-emitting diodes (OLEDs) with high color purity have emerged as promising alternatives. Traditional OLEDs, typically based on a bottom-emitting structure (BE-OLEDs), face limitations in terms of color purity and light extraction efficiency, making them less suitable for ultrahigh-definition displays (5). In contrast, top-emitting OLEDs (TE-OLEDs) can effectively narrow the emission spectrum and enhance the color purity through microcavity effects while also mitigating light absorption through underlying layers such as thin-film transistor backplanes, thereby improving the light extraction efficiency and operational lifetimes (6–9). These advantages have attracted notable industrial interest in TE-OLEDs, making them highly suitable for applications requiring a high color accuracy and wide color gamut, and they have already entered the commercialization phase (10). Owing to structural deformations and molecular vibrations in the excited state, organic materials typically exhibit broad emission spectra accompanied by vibrational shoulders (11–13), which increase optical losses caused by microcavities, thereby reducing their efficiency and stability. Therefore, the development of organic emitters with intrinsically narrow emission spectra is urgently needed. Notably, the human visual system exhibits particular sensitivity in the green wavelength region; thus, even slight spectral broadening or shoulder formation effects, leading to color mixing with adjacent blue/red bands, are readily perceptible. Therefore, achieving high-purity green emission is a more stringent and challenging target (14). Recently, some multiple-resonance thermally activated delayed fluorescence (MR-TADF) green emitters have demonstrated impressive color purity (15, 16). For instance, TCz-B (17), TRZCzPh-BNCz (18), DBTN-2 (19), and ω-DABNA (20) exhibit electroluminescence (EL) spectra with FWHM values of 30, 33, 29, and 25 nm, respectively, and corresponding Commission Internationale de l’Éclairage (CIE) y coordinates of 0.71, 0.70, 0.74, and 0.73, closely aligning with the CIEy coordinate requirement for the green color of the BT.2020 standard. Similarly, rigid tetradentate Pt(II) complexes have shown excellent color purity, with PtN7N (21) exhibiting FWHM values of 24 and 20 nm in BE-OLEDs and TE-OLEDs, corresponding to CIEy coordinates of 0.67 and 0.74, respectively. Despite these substantial achievements in terms of color purity, the operational lifetime remains a critical challenge for commercial ultrahigh-definition display applications. Current strategies using stable phosphorescent or TADF sensitizers as triplet-harvesting assistants have successfully extended the operational lifetime of MR-TADF–based devices (22–26). Furthermore, the use of TADF or exciplex hosts has shown promising potential in enhancing the lifetime of phosphorescent OLEDs (27). Despite these improvements, the operational lifetimes of reported green OLEDs still fall substantially short of those achieved by commercial micro–light-emitting diodes (MicroLEDs). According to LG’s official specifications for the MAGNIT LAAA015 MicroLED display, this device achieves a half-brightness lifetime of ∼100,000 hours when driven at a peak brightness of around 1000 nits. Consequently, the development of green OLEDs that simultaneously satisfy the demanding requirements of high color purity, high efficiency, and long operational lifetime continues to represent a formidable challenge in the field of display technologies. Addressing this multifaceted challenge necessitates the development of innovative material designs and device architectures. RESULTS Molecular design and theoretical calculations Tetradentate Pt(II) complexes typically exhibit emission spectra with FWHM values exceeding 20 nm and pronounced v0-1 vibrational sidebands, which substantially reduce their color purity (28–34). Therefore, further narrowing the intrinsic emission spectra of Pt(II) complexes and reducing vibrational sidebands are crucial. The emission spectra of Pt(II) complexes are mainly determined by the nature of the lowest triplet excited state (T1), which typically comprises a mixture of locally excited (3LE), metal-to-ligand charge transfer (3MLCT), and intraligand charge transfer (3ILCT) characteristics (Fig. 1A) (31–33). Among these, the 3LE state is crucial for achieving a narrow emission spectrum, owing to its minimal geometric distortion upon excitation, whereas the 3MLCT component facilitates efficient intersystem crossing (ISC) of S1 → T1 → S0 states, thereby enhancing the quantum efficiency. To simultaneously achieve a narrow emission bandwidth and high efficiency, it is essential to fine-tune the excited state in such a way that the 3LE energy level is located slightly below the 3MLCT state, resulting in a 3LE-dominated T1 state mixed with a 3MLCT character (Fig. 1A). Furthermore, Pt(II) complexes often adopt a square-planar geometry, which causes them to form intermolecular π-π stacking and Pt-Pt interactions. Such interactions can lead to broadened and red-shifted emission spectra, as well as to negative effects such as triplet-triplet annihilation (TTA) and efficiency roll-off in devices. Substitution of steric hindrance groups can help to avoid close packing and aggregation, thereby preserving the spectral purity and device performance. On the basis of these considerations, we designed two robust emitters, PtN5N1 and PtN5N2 (Fig. 1B). To enhance the 3LE character of the T1 state, we incorporated an extended π-conjugated carbazole-carbazole (Cz-Cz) moiety, forming a rigid fused 5/6/6 metallocyclic structure with pyridine (Py) and benzocarbene (Ph/NHC) ligands. We then introduced a bulky 3,5-di-tert-butylphenyl (DTP) group on the Ph/NHC moiety to obtain PtN5N1 and a more sterically demanding 3,3″,5,5′,5″-penta-tert-butyl-1,1′:3′,1″-terphenyl (PTBTP) group to form PtN5N2. The larger PTBTP group would enforce a three-dimensional (3D) molecular geometry in PtN5N2, effectively suppressing intermolecular interactions. Moreover, these sterically hindered substituents could restrict rotation around the CPh─NNHC bond that substantially enhanced molecular rigidity. This increased structural rigidity would not only reduce nonradiative decay pathways by limiting vibrational and rotational motions but also reduce the Huang-Rhys factor (HRF; SM) and results in enhanced color purity. Molecular design and theoretical calculations. (A) Potential energy curves between S0 and T1 states, illustrating the design of narrowband Pt(II) complexes via regulation of their excited-state properties. (B) Molecular structures of PtN5N1 and PtN5N2. (C) DFT-calculated frontier molecular orbital distributions, energy levels, and energy bandgaps (Eg) of PtN5N1 and PtN5N2. (D) NTO analyses of T1 states for PtN5N1 and PtN5N2. Ph/NHC, phenyl N-heterocyclic carbene; Py, pyridine; Cz, carbazole. Density functional theory (DFT) calculations (Fig. 1C) revealed that PtN5N2 had slightly shallower highest occupied molecular orbital (HOMO; −4.39 eV) and lowest unoccupied molecular orbital (LUMO; −1.25 eV) levels compared to PtN5N1 (HOMO, −4.42 eV; LUMO, −1.28 eV) while maintaining an identical energy gap (Eg) of 3.14 eV. This suggests that the PTBTP group did not substantially alter the emission characteristics. The two complexes exhibited similar frontier orbital distributions, with the HOMO localized on the Cz1-Cz2-π and Pt-d orbitals and the LUMO primarily concentrated on the pyridine-π and benzocarbene-π systems. Time-dependent DFT (TD-DFT) and natural transition orbital (NTO) analyses of the T1 state (Fig. 1D) indicated that both PtN5N1 and PtN5N2 had 3LE characters predominantly localized on the Cz1-Cz2 units, with 43.5 and 43.1% contributions from πCz1-Cz2* orbitals, respectively. In addition, the 3MLCT components of PtN5N1 and PtN5N2 were estimated to be 7.8 and 9.2%, respectively, based on the decreased electron density on the Pt atom (35). These findings confirm that the T1 states of both complexes had a 3LE-dominated character mixed with large 3MLCT contributions, consistent with the desired narrow-spectrum emission. Synthesis and characterization The synthetic routes for PtN5N1 and PtN5N2 are illustrated in Fig. 2A. Initially, a copper(I)-catalyzed Ullmann coupling reaction was used to couple 2-bromo-9-[4-(tert-butyl)pyridin-2-yl]-9H-carbazole (Py-Cz) with 2-chloro-9H-carbazole (Cz-Cl), yielding 9-[4-(tert-butyl)pyridin-2-yl]-2′-chloro-9H-2,9′-bicarbazole (Py-CzCz) with an 86% yield. Subsequently, Py-CzCz underwent Pd(0)-catalyzed C–N cross-coupling reactions with TBBI-NH2 and PTBTP-NH2 (34) to form electron-rich diarylamine intermediates, each obtained in 84% yield. Because of their susceptibility to oxidation, these intermediates were promptly purified and subjected to cyclization reactions with triethyl orthoformate, affording the benzimidazole ligand precursors LN5N1 and LN5N2 in yields of 99 and 80%, respectively. Last, metalation of LN5N1 and LN5N2 with PtCl2 and Pt(PhCN)2Cl2, respectively, produced the corresponding tetradentate Pt(II) complexes PtN5N1 and PtN5N2 in yields of 30 and 39%. All intermediates, ligands, and Pt(II) complexes were characterized by 1H and 13C nuclear magnetic resonance (NMR) spectroscopy, high-resolution mass spectrometry (HRMS), as well as high-performance liquid chromatography (HPLC) with detailed spectral data provided in the Supplementary Materials (figs. S1 to S30). Synthesis and single-crystal x-ray diffraction structures. (A) Synthetic route of PtN5N1 and PtN5N2. h, hours; XPhos, 2-dicyclohexylphosphino-2′,4′,6′-triisopropylbiphenyl; DMF, N,N′-dimethylformamide; o-DCB, 1,2-dichlorobenzene. (B and C) ORTEP (Oak Ridge thermal ellipsoid plot) drawings of single-crystal x-ray diffraction molecular structures of PtN5N1 and PtN5N2; hydrogen atoms are omitted for clarity. The single-crystal structures of PtN5N1 and PtN5N2, as analyzed by x-ray diffraction, are illustrated in Fig. 2 (B and C) and table S1. In PtN5N1, the dihedral angle between the phenyl group and the Ph/NHC plane (α) was 50.88°, whereas that between the terminal pyridine and the Ph/NHC plane (∠A/B) was 50.62°; these angles were 65.88° and 56.49° in PtN5N2, indicating a more twisted 3D geometry attributed to the bulky PTBTP substituent. The increased steric hindrance in PtN5N2 enhances the molecular rigidity and mitigates intermolecular interactions. PtN5N1 exhibited notable π-π stacking interactions in its crystal packing, whereas such interactions were absent in PtN5N2 (figs. S31 and S32). Despite the twisted molecular structures, the sums of the four coordination angles around Pt(II) ions in PtN5N1 and PtN5N2 were 361.90° and 361.91° (table S2), respectively, closely approximating the ideal 360° value for a square-planar geometry, consistent with the dsp2 hybridization characteristic of the Pt(II) centers. Photophysical properties and thermal stability The absorption spectra of PtN5N1 and PtN5N2 are shown in Fig. 3A, whereas their photophysical properties are summarized in table S3. On the basis of NTO analysis of the S1 state (fig. S33), the intense absorption bands below 400 nm (ε > 1.0 × 104 cm−1 M−1) were attributed to ligand-centered π-π* transitions, primarily involving πCz-Cz → πCz-Cz*, πCz1 → πPh/NHC*, and πCz2 → πPy* processes. The absorption bands in the 400- to 500-nm range (ε > 2.4 × 103 cm−1 M−1) were assigned to spin-allowed 1MLCT transitions, such as πCz1dPt → πNHC* and πCz2dPt → πPy*. In addition, weak S0 → T1 absorptions (ε < 100 cm−1 M−1) were observed in the 505- to 535-nm region, corresponding to spin-flip 3MLCT transitions from dPt to πPh/NHC* and dPt to πPy*, consistent with the T1 state energies determined at 77 K. Both PtN5N1 and PtN5N2 exhibited strong narrowband green phosphorescence (Fig. 3, B and C). In 2-methyltetrahydrofuran (2-MeTHF) at 77 K, their emission peak was located at 515.4 and 521.2 nm, corresponding to triplet energy levels (ET1) of 2.41 and 2.38 eV, respectively. At room temperature in toluene, the emission peaks exhibited red shifts to 530.1 and 531.0 nm, with FWHM values of 20.4 and 16.4 nm, along with SM values of 0.283 and 0.248, respectively. The SM values approximated the height ratio of the v0-1 to v0-0 peaks (36). The reduced red shift, smaller FWHM, and lower SM values observed for PtN5N2 are attributed to the enhanced molecular rigidity imparted by the PTBTP group, which suppresses excited-state structural relaxation and improves the color purity. Notably, compared to the 77 K spectra, an additional emission band appeared in the shorter wavelength range (475 to 510 nm) at room temperature. This is attributed to metal-assisted delayed fluorescence (MADF) (37, 38), arising from thermal upconversion of T1 to S1 through reverse ISC (39–41). Photophysical properties. (A) Absorption spectra of LN5N1, LN5N2, PtN5N1, and PtN5N2 at room temperature in dichloromethane. (B and C) PL spectra of PtN5N1 and PtN5N2 at 77 K in 2-MeTHF and at room temperature in toluene. a.u., arbitrary units. (D) Relationship between FWHM and SM values of PtN5N1 in various solvents. (E) Theoretically calculated HRFs for T1 → S0 transitions in PtN5N1 and PtN5N2. (F) Transient decay curves of 10 wt % Pt(II) emitter-doped in premixed host films. The Φ and τ of each film are shown in the inset. (G) TGA curves of PtN5N1 and PtN5N2. ΔT5% denotes the decomposition temperature at 5% weight loss. Solvatochromic analysis revealed that increasing the solvent polarity induced a slight blue shift in the emission spectra of both complexes (figs. S34 and S35). This behavior arises from their smaller dipole moments in the T1 state (5.27 D for PtN5N1 and 4.78 D for PtN5N2) than the S0 state (9.18 and 8.44 D, respectively), as determined by DFT calculations. Across solvents ranging from nonpolar hexane to highly polar acetonitrile, the emission peak shifted by only 8.6 nm for PtN5N1 and 8.3 nm for PtN5N2, indicating minimal solvatochromic effects. Furthermore, both complexes consistently exhibited narrow emission spectra, with FWHM values of 16.4 to 24.4 and 13.2 to 17.6 nm, as well as SM values in the 0.234 to 0.315 and 0.194 to 0.280 ranges for PtN5N1 and PtN5N2, respectively. The separations between the 0-0 and 0-1 vibrational bands, within 1216 to 1279 cm−1 for PtN5N1 and 1255 to 1327 cm−1 for PtN5N2, are representative of stretching vibrational modes (tables S4 and S5). A strong linear correlation between FWHM and SM values was observed for PtN5N1 across all solvents (Fig. 3D) and for PtN5N2 in all solvents except tert-butylbenzene, ethanol, and methanol (fig. S36). These findings suggest that the T1 states of both complexes exhibited a predominantly 3LE character. To further elucidate the narrowband emission mechanisms, we calculated the HRFs and reorganization energies for the T1 → S0 transitions of the two complexes (Fig. 3E and fig. S37). Despite the introduction of additional twisting vibrations by the PTBTP group in PtN5N2, this compound exhibited lower HRF values and reorganization energies in the frequency range of 0 to 1000 cm−1 compared to PtN5N1. Crucially, the intrinsic narrow spectra of PtN5N1 and PtN5N2 primarily originate from the 3LE character, and thus the vibrational modes of the Cz1-Cz2 moiety notably affect the emission spectra. In the high-frequency region (>1000 cm−1), PtN5N2 showed reduced stretching/scissoring vibrations involving the Cz1-Cz2 moiety (figs. S38 and S39). For instance, the HRF values measured for PtN5N2 at 1312.23, 1307.30, 1304.60, and 1287.38 cm−1 (0.072, 0.012, 0.015, and 0.096, respectively) were lower than those obtained for PtN5N1 at 1315.05, 1309.54, 1304.82, and 1287.84 cm−1 (0.082, 0.060, 0.042, and 0.113, respectively). The corresponding reorganization energies were also reduced in PtN5N2. The smaller HRFs and reorganization energies for PtN5N2 are consistent with its reduced FWHM and SM values. Transient photoluminescence decay measurements in 5 wt % doped polymethyl methacrylate (PMMA) films revealed excited-state lifetimes (τ) of 8.21 and 7.92 μs for PtN5N1 and PtN5N2, respectively (fig. S40). The slightly shorter τ for PtN5N2 is attributed to an increased 3MLCT character. Consistently, PtN5N2 exhibited a slightly higher photoluminescence quantum yield (ΦPL) of 90% compared to that (86%) of PtN5N1. When investigated in 10 wt % doped evaporated host films, PtN5N1 and PtN5N2 exhibit τ values of 6.59 and 7.09 μs, respectively (Fig. 3F), together with remarkably high ΦPL values of 94 and 99%. Thermogravimetric analysis (TGA) revealed that PtN5N1 and PtN5N2 had exceptional thermal stability, with decomposition temperatures at 5% weight loss (ΔT5%) of 525° and 466°C, respectively (Fig. 3G). Such high thermal stability is advantageous for subsequent sublimation purification and for the fabrication of OLED devices through thermal evaporation because it minimizes the risk of thermal decomposition during processing. Device performances On the basis of these favorable photophysical and thermal characteristics, OLED devices using PtN5N1 and PtN5N2 as emitters were subsequently fabricated and systematically investigated. First, BE-OLEDs were fabricated with the following device architecture: indium tin oxide (ITO)/BFSFA:FCN (97:3, 10 nm)/BFSFA (130 nm)/BFSFAtBu (50 nm)/premixed host:emitter (10 wt %, 30 nm)/NAPPBI:Liq (50:50, 35 nm)/Yb (1 nm)/Ag (80 nm) (Fig. 4A), where BFSFA = N,N-bis(9,9-dimethyl-9H-fluoren-2-yl)-9,9′-spirobi[fluoren]-4-amine, FCN = 4,4′,4″-((1E,1′E,1″E)-cyclopropane-1,2,3-triylidenetris(cyanomethaneylylidene))tris(2,3,5,6-tetrafluorobenzonitrile), BFSFAtBu = 2′,7′-di-tert-butyl-N,N-bis(9,9-dimethyl-9H-fluoren-2-yl)-9,9′-spirobi[fluoren]-4-amine, NAPPBI = 1-(4-(10-(naphthalen-2-yl)anthracen-9-yl)phenyl)-2-phenyl-1H-benzo[d]imidazole and Liq = lithium quinolin-8-olate. The chemical structures of the organic layers are provided in fig. S41, and a photograph of the PtN5N2-based device is displayed in Fig. 4B. The BE-OLEDs based on PtN5N1 and PtN5N2 exhibited outstanding electroluminescent performances, with low operating voltages of 2.98 and 3.07 V at a high luminance of 1000 cd/m2, respectively (Fig. 4C). Both devices exhibited pure green emission with peaks at 533 nm (PtN5N1) and 532 nm (PtN5N2), accompanied by remarkably small FWHM values of 23.7 and 20.3 nm, respectively (Fig. 4D). The maximum external quantum efficiency (EQEmax) reached 17.1% for PtN5N1 and 19.0% for PtN5N2. Notably, at a practical luminance of 1000 cd/m2, the EQEs remained as high as 16.1 and 18.0%, respectively, indicating minimal efficiency roll-off in the practical operating regime (Fig. 4E). These values are higher than those of the Ir(ppy)3-based reference devices fabricated with the same device architecture, which exhibit EQEs of 15.3% at the maximum and 15.1% at 1000 cd/m2 (fig. S42). The enhanced EQE of the PtN5N1- and PtN5N2-based devices can be partially attributed to their favorable horizontal dipole orientation, with dipole orientation ratios (Θ) of 76 and 74%, respectively, whereas the Ir(ppy)3-based device shows a more isotropic orientation with a lower Θ value of 67% (fig. S43). Moreover, the operational stability of the devices was evaluated at a constant current density of 50 mA/cm2, yielding LT90 (defined as the time at which the luminance decays to 90% of its initial level) values of 136 and 58 hours for PtN5N1 and PtN5N2, respectively (Fig. 4F). Properties of BE-OLEDs. (A) Schematic illustration of the BE-OLED device structure. (B) Lighting of the ZJUT logo by the PtN5N2-doped device. (C) Current density–voltage–luminance (J-V-L) curves. (D) EL spectra of 10 wt % PtN5N1- and PtN5N2-doped green OLEDs at 50 mA/cm2. (E) External quantum efficiency versus luminance (EQE-L) curves. (F) Operational lifetimes of pure green OLEDs at 50 mA/cm2. Because commercial OLED displays adopt top-emitting configurations, and both PtN5N1 and PtN5N2 exhibit intrinsically narrowband PL and EL spectra, we constructed a top-emitting device structure with microcavity effect as follows: ITO/Ag/ITO (12 nm/100 nm/10 nm)/BFSFA:FCN (97:3, 10 nm)/BFSFA (145 nm)/BFSFAtBu (50 nm)/premixed host:emitter (10 wt %, 30 nm)/NAPPBI:Liq (50:50, 35 nm)/Yb (1 nm)/Ag:Mg (90:10, 14 nm)/BPNBNBA (70 nm) (Fig. 5A and fig. S41). BPNBNBA (N-(4-(benzo[d]oxazol-2-yl)phenyl)-N-(naphtho[2,3-b]benzofuran-3-yl)naphtho[2,3-b]benzofuran-3-amine) serves as an efficient capping layer (CPL) on the top cathode to enhance light extraction in TE-OLEDs. Angle-dependent EL measurements were performed to further elucidate the optical characteristics of the top-emitting microcavity devices. As shown in fig. S44, the BE OLEDs exhibit a near-Lambertian emission profile, whereas the TE devices display a pronounced non-Lambertian angular distribution, directly confirming the strong microcavity modulation in the TE architecture and its critical role in shaping the emission profile and spectral characteristics of the devices. Through the Fabry-Pérot resonance effect, constructive interference is achieved at the emission peak while destructive interference effectively suppresses the vibronic sidebands (Fig. 5B). This optical engineering approach simultaneously enhances the peak emission intensity while reducing the FWHM, resulting in substantially improved color purity. Benefiting from the microcavity effect, PtN5N1- and PtN5N2-based TE-OLEDs exhibited low operating voltages of 2.76 and 2.74 V at 1000 cd/m2, respectively. Even at 10,000 cd/m2, the voltages remained as low as 3.64 and 3.53 V, respectively; these values were notably lower than those of BE-OLEDs (Fig. 5C). Notably, PtN5N2 exhibited an intrinsically smaller FWHM (16.4 nm in solution versus 20.4 nm for PtN5N1), owing to its more rigid 3D molecular structure. This inherent spectral advantage is further amplified by the microcavity effect, resulting in a higher luminance at identical voltages (Fig. 5C). As shown in Fig. 5D, the emission peaks of PtN5N1- and PtN5N2-based TE-OLEDs were located at 533 and 532 nm, respectively, with even smaller FWHM values of 16.6 and 14.8 nm, which are smaller than those of MicroLEDs and QD-LEDs (42–44). The corresponding CIE coordinates were (0.23, 0.74) and (0.22, 0.75), respectively, resulting in ultrapure green emission. These color coordinates exceed the green standard set by the National Television Standards Committee (NTSC) (CIEy = 0.71) and are very close to the BT.2020 green standard specifications (Fig. 5E). Moreover, small angular color shifts are observed for the PtN5N1- and PtN5N2-based TE-OLEDs, with peak shifts limited to 4.5 and 3.7 nm from 0° to 60° (fig. S45), respectively, which are much smaller than that of the Ir(ppy)3-based reference device (27.9 nm), suggesting suppressed microcavity-induced angular color dispersion. Properties of TE-OLEDs. (A) Schematic illustration of TE-OLEDs with microcavity structure. (B) Microcavity effect diagram; the microcavity induces constructive interference at the main peak, leading to increased amplitude and enhanced light intensity, while causing destructive interference at the shoulder peak, resulting in reduced amplitude and decreased intensity. (C) J-V-L curves. (D) EL spectra of 10 wt % PtN5N1- and PtN5N2-doped TE-OLEDs at 50 mA/cm2. (E) Corresponding CIE coordinates. (F) Current efficiency–luminance–power efficiency (CE-L-PE) curves. (G and H) Statistical CE data of PtN5N1 (22 devices) and PtN5N2 (24 devices), respectively. (I and J) Operational lifetimes of TE-OLEDs at 30, 40, and 50 mA/cm2 for PtN5N1 and PtN5N2, respectively. (K) Degradation acceleration factor (n) of TE-OLEDs. (L) 3D plot showing LT95 @ 1000 cd/m2, CIEy, and FWHM values of representative green OLEDs (see table S6 for details). The devices also exhibited outstanding efficiency, with peak current efficiencies (CEs) of 161 and 174 cd/A and power efficiencies (PEs) of 186 and 202 lm/W for PtN5N1- and PtN5N2-based TE-OLEDs, respectively (Fig. 5F). Under high current density operation, the current density–dependent EL characteristics reveal a more pronounced efficiency roll-off for the PtN5N1- and PtN5N2-based devices compared with the Ir(ppy)3-based reference devices (fig. S46). Such roll-off behavior in phosphorescent OLEDs is commonly associated with exciton density–dependent loss processes, including TTA and triplet-polaron quenching (TPQ) (45). Consistent with this interpretation, both transient photoluminescence (fig. S40) and transient EL measurements (fig. S47) indicate that PtN5N1 and PtN5N2 have substantially longer excited-state lifetimes than Ir(ppy)3 in both doped films and operating devices. In particular, the transient EL results reveal exciton lifetimes of 4.28 and 4.45 μs for the PtN5N1- and PtN5N2-based TE devices, respectively, which are notably longer than that of the Ir(ppy)3-based device (1.25 μs). The relatively long-lived triplet states lead to a higher steady-state triplet exciton density under high current operation, which enhances nonradiative exciton dissipation through TTA and TPQ pathways and consequently results in a more pronounced efficiency roll-off at extremely high current densities (45). To evaluate the reliability and reproducibility of the device performance, a statistical analysis of the EL efficiencies was carried out using multiple devices (Fig. 5, G and H, and figs. S48 and S49). The results show that both PtN5N1- and PtN5N2-based devices exhibit only small device-to-device variations, demonstrating good reproducibility of the fabrication process. Under the same constant current density of 50 mA/cm2, the LT90 lifetimes of PtN5N1 and PtN5N2 devices with initial luminance (L0) values of 41,827 and 44,634 cd/m2 were 165 and 88 hours, respectively (Fig. 5, I and J), which are substantially longer than those of the corresponding BE-OLEDs. This improvement is mainly attributed to the reduced optical and energy losses in the top-emitting device configuration. The LT values at different luminance levels can be predicted using the accelerated degradation model, which follows the empirical relationship L0n · LT = constant (46). This relationship can be equivalently expressed as LT(L1) = LT(L0) · (L0/L1)n, where L1 represents the target luminance and n denotes the degradation acceleration factor. As illustrated in Fig. 5K, the n values for the PtN5N1 and PtN5N2 devices were calculated as 2.17 and 2.14, respectively, by fitting the LT90-luminance dependence using experimental lifetime data obtained at 50, 40, and 30 mA/cm2. Repeated lifetime measurements further confirm the robustness of the acceleration factors, yielding values of 2.20 and 2.21 for PtN5N1 and PtN5N2 (fig. S50), respectively, which differ by less than 0.1 from those obtained above. In addition, the high device-to-device consistency of the EL characteristics indirectly supports the reliability of the extracted lifetime parameters. For a conservative evaluation, the lifetime extrapolations are based on the smaller acceleration factors. The relatively high acceleration factors observed for the PtN5N1- and PtN5N2-based devices reflect the intrinsic emitter properties and the dominant degradation mechanism involved. Consistent with a previous report, the n value is not a universal constant but depends on both the nature of the emitter and the exciton loss processes governing device aging (46). Control experiments using Ir(ppy)3-based reference devices fabricated with the same device architecture yield an n value of 1.67, comparable to typical Ir-based phosphorescent OLEDs (typically n ≈ 1.7) (47), indicating that the larger n values observed here primarily originate from the intrinsic characteristics of the Pt(II) emitters. In particular, the intrinsically long-lived triplet states promote enhanced exciton density–dependent annihilation processes, such as TTA and TPQ (45), which is consistent with the more pronounced efficiency roll-off in the Pt(II)-based devices (fig. S46). As a result, these nonlinear exciton loss processes account for the larger n values observed for the Pt(II)-based devices. The estimated LT90 lifetimes of the PtN5N1- and PtN5N2-based devices remain 3651 and 2168 hours at an initial luminance of 10,000 cd/m2, respectively, and increase to 540,100 and 299,200 hours at 1,000 cd/m2 (Table 1). Under the same extrapolation conditions, the Ir(ppy)3-based device exhibits an LT90 lifetime of ∼44,127 hours at 1000 cd/m2 (fig. S51). In comparison, the PtN5N1- and PtN5N2-based devices show markedly extended LT90 lifetimes, corresponding to a maximum enhancement of about 12.2 times. For comparison with the reported data, we also extrapolated the LT95 lifetimes at 1000 cd/m2 to be 279,100 and 154,700 hours, respectively (Table 1). To the best of our knowledge, PtN5N1 represents the most stable green OLED reported to date, exhibiting a device lifetime more than ∼59 times longer than the longest lifetime reported for pure green OLEDs with CIEy ≥ 0.70 (Fig. 5L and table S6). The operational lifetimes of our PtN5N1 and PtN5N2 TE-OLEDs far surpass those of reported ultrastable green perovskite LEDs (PeLEDs) and QD-LEDs (table S7), as well as greatly exceed standard commercial benchmarks for LED, OLED, and MicroLED displays (table S8). The exceptional operational stability of the PtN5N1- and PtN5N2-based devices originates from the combined contributions of the high intrinsic stability of the emitters and the optimized top-emitting device architecture. The excellent electrochemical (fig. S52) and thermal stability of the Pt(II) emitters provides a robust intrinsic foundation for long device lifetime. Consistent with this interpretation, reference devices based on Ir(ppy)3 exhibit notably shorter operational lifetimes than those based on PtN5N1 and PtN5N2, indicating that the superior intrinsic stability of the Pt(II) emitters plays a decisive role. In addition, their inherently narrow emission bandwidth allows more efficient photon utilization in the microcavity, thereby mitigating spectral truncation losses (fig. S53) and enabling high luminance to be achieved at lower driving voltages. The reduced operating voltage under identical luminance conditions effectively suppresses degradation processes. Moreover, the substantially extended operational lifetime of Ir(ppy)3-based reference devices compared to previously reported values (48) further confirms the robustness of the used host material system. Together with the optimized top-emitting device architecture, this well-established host material system contributes to the highly extended operational lifetimes observed in these devices. It is worth mentioning that the CIE coordinates, efficiency, and lifetimes of the PtN5N1 and PtN5N2 devices also meet the stringent requirements of commercial wide-color-gamut green OLED displays (1, 19, 49–51). Table 1.Summary of device performance data.DevicePtN5N1(BE)PtN5N2(BE)PtN5N1(TE)PtN5N2(TE)λmax (nm)533532533532FWHM (nm)23.720.316.614.8CIE coordinates(0.33, 0.65)(0.31, 0.65)(0.23, 0.74)(0.22, 0.75)EQEmax (%)17.119.0––EQE (%) @1000 cd/m216.118.0––CEmax (cd/A)67.475.6161174CE (cd/A) @1000 cd/m263.271.7159173CE (cd/A) @10,000 cd/m243.250.7134139L0 (cd/m2) @50/40/30 mA/cm217,593/–/–19,765/–/–41,827/36,029/29,52844,634/38,213/31,042LT95 (hours) @L069/–/–28/–/–86/115/18045/64/100LT95 (hours) @10,000 cd/m2––18871121LT95 (hours) @1000 cd/m2––279,100154,700LT90 (hours) @L0136/–/–58/–/–165/222/35288/124/191LT90 (hours) @10,000 cd/m2––36512168LT90 (hours) @1000 cd/m2––540,100299,200 Summary of device performance data. DISCUSSION In summary, we have developed two pure green phosphorescent emitters, PtN5N1 and PtN5N2, based on tetradentate Pt(II) complexes. By extending the π-conjugation through a Cz1-Cz2 moiety design, these emitters exhibited 3LE-dominated emission; in addition, incorporating a bulky PTBTP group enabled the fabrication of a rigid 3D geometry to suppress intermolecular interactions. This structural strategy notably reduced geometric deformation in the excited state, resulting in exceptionally small FWHM values of 20.4 and 16.4 nm in toluene for PtN5N1 and PtN5N2, respectively. Benefiting from their intrinsic narrowband emission, microcavity-enhanced TE-OLEDs fabricated with these emitters achieved further reduced FWHM values of 16.6 and 14.8 nm, along with CIE coordinates of (0.23, 0.74) and (0.22, 0.75), respectively, indicating ultrapure green emission. Most notably, the acceleration factors for the PtN5N1 and PtN5N2 devices were calculated as 2.17 and 2.14, respectively. Using these n values, the operational lifetimes were extrapolated, yielding estimated LT90 lifetimes of 540,100 hours for PtN5N1 and 299,200 hours for PtN5N2 at 1000 cd/m2. These values substantially exceeded that of the standard commercial benchmarks for LED, OLED, and MicroLED displays. These CIE coordinates, efficiency, and lifetimes meet the stringent requirements for commercial wide-color-gamut green OLED displays. This work demonstrates the feasibility of preparing stable, high-efficiency, and high-color-purity green OLEDs; in addition, it provides a valuable molecular design strategy for developing phosphorescent emitters with narrowband emission. MATERIALS AND METHODS Synthesis and structural characterization Unless noted, all commercial reagents were purchased and used as received without further purification. 1H NMR spectra were recorded at 400 or 500 MHz, and 13C NMR spectra were recorded at 125-MHz NMR Bruker instruments. All of the new compounds were analyzed for HRMS on a Waters mass spectrometer using electrospray ionization in positive ion mode of ESI-Q-TOF (electrospray ionization–quadrupole orthogonal acceleration–time-of-flight). X-ray crystallography measurements X-ray diffraction data were collected at 170 K on a Bruker D8 Venture diffractometer, using graphite-monochromated Mo Kα radiation (λ = 0.71073 Å) from a rotating anode generator. Quantum chemical calculations The theoretical calculations were performed using the Gaussian 16 package. The molecular geometries of the S0 and T1 states were optimized with the DFT method at the B3LYP level. A 6-31G(d) basis set was used for C, H, and N atoms, whereas a LANL2DZ basis set was used for the Pt atom. The energies of the singlet and triplet excited states of the complexes were calculated using the TD-B3LYP method with SV, SVP, and def2-TZVP basis sets for H, C/N, and Pt atoms, respectively, based on the optimized S0 geometry. Photophysical measurements Absorption spectra were measured on a Hitachi U-3900 UV-Vis spectrometer. Steady-state emission experiments were performed on a Hitachi F-7000 spectrometer. Low-temperature (77 K) emission spectra and lifetimes were measured in 2-MeTHF cooled with liquid nitrogen. Lifetime and quantum efficiency measurements were performed using an Edinburgh FS5 spectrofluorometer equipped with an integrating sphere. OLED fabrication and characterization BE-OLEDs and TE-OLEDs were fabricated and tested in an R&D evaporation system (OASIS-C-E4T1-PcG-2020, Choshu Industry Co. Ltd.). The current density–voltage–luminance characteristics of the OLEDs were measured using a Keithley 2400 source meter and a Konica Minolta CS-2000A source multimeter equipped with a calibrated silicon photodiode, following standard procedures (52). The areas of the devices were 6.25 mm2 (2.5 mm by 2.5 mm).