Blue light emission is paramount in light-emitting devices due to its high energy among primary colors in displays and its role in generating white light for lighting applications. While OLEDs have been successfully commercialized for various applications, including smartphones and large-screen displays, owing to their ability to produce high-color images with large contrast, blue OLEDs suffer from the drawback of requiring a high applied voltage. This is because the energy of blue emission is approximately 3 eV, leading to typical blue OLEDs needing around 4 V for a luminance of 100 cd/m², a standard display condition. The industry aims to operate blue OLEDs within 3.7 V, aligning with the voltage of lithium-ion batteries in most mobile devices. Therefore, low-voltage blue OLEDs are highly sought after to meet commercial demands. Conventional fluorescent emitters remain prevalent in commercial blue OLEDs due to their reliability and long operational lifetime, despite lower external quantum efficiencies (EQEs) compared to phosphorescent and thermally activated delayed fluorescence (TADF) materials. The challenge with phosphorescent and TADF materials lies in their high-energy first triplet excited state (T1), often close to 3 eV, which is needed for their operational mechanism. This high energy level is problematic because it is comparable to the dissociation energy of a carbon-nitrogen bond, leading to material degradation. While recent advancements have yielded phosphorescent and TADF materials with improved stability, their intrinsic limitations hinder their widespread commercial adoption in blue OLEDs. This research aims to address this challenge by exploring alternative approaches to achieve efficient blue emission at low voltages using conventional fluorescent emitters.

Previous research on upconversion (UC) OLEDs, primarily using rubrene (which emits yellow light), has demonstrated the potential of using the charge transfer (CT) state at the donor-acceptor (D/A) interface as an energy transfer pathway to the triplet states of the emitter. Triplet-triplet annihilation (TTA) is then used to generate a singlet state which results in light emission. However, achieving this efficiently with a blue emitter had not been successfully demonstrated prior to this research. The energy of the first triplet excited state (T1) of anthracene derivatives, common fluorescent emitters used in blue OLEDs, is typically around 1.7 eV, considerably lower than the energy required for direct blue emission. The study investigates whether selective excitation of this low-energy T1 followed by TTA can lead to efficient blue emission at significantly reduced voltages.

The researchers used 1,2-ADN (9-(naphthalen-1-yl)-10-(naphthalen-2-yl)anthracene), a widely used host material in blue fluorescent OLEDs, as the emitter (donor) in their UC-OLEDs. They investigated four different acceptor (electron transport) materials: two phenyl pyridine derivatives (TmPyPB and B4PYMPM), a bipyridyl substituted oxadiazole derivative (BPyOXD), and a naphthalene diimide derivative with a fluorene side chain (NDI-HF). The energy levels of these materials were evaluated using photoelectron spectroscopy and absorption spectra. Bilayer-type OLED devices were fabricated using thermal evaporation under high vacuum, with the selected emitter and acceptor materials layered with the respective hole and electron transport layers. OLED device characteristics such as current density, luminance, and EQE were analyzed. Time-resolved electroluminescence measurements provided insights into the emission decay dynamics, distinguishing between prompt fluorescence from the singlet state (S1) and delayed fluorescence originating from TTA. Incident photon-to-current conversion efficiency (IPCE) was used to study CT state absorption. To further optimize the device performance, 14 NDI derivatives with various substituents were synthesized and investigated as electron-transporting materials to identify the optimal acceptor for efficient energy transfer and TTA-UC emission. Finally, tert-butyl perylene (TbPe) was employed as a dopant in the emitter layer to optimize device performance.

The study revealed a significant difference in the turn-on voltage of OLEDs based on the choice of the acceptor material. Devices with NDI-HF as the acceptor demonstrated a drastically lower turn-on voltage (1.7 V) compared to those with other acceptor materials (2.8-3.5 V). The electroluminescence (EL) decay analysis showed a slow decay component in the 1,2-ADN/NDI-HF device, indicative of TTA-UC emission, while other devices exhibited mostly prompt decay from S1. IPCE measurements confirmed the formation of a CT state in the 1,2-ADN/NDI-HF device, but not in the devices with other acceptor materials. This CT state acts as the precursor for selective excitation of the low-energy T1 state. An investigation of 14 NDI derivatives highlighted a negative correlation between the CT state energy and TTA-UC emission intensity, with the lowest CT state energy found in the 1,2-ADN/NDI-HF device, resulting in the most efficient TTA-UC emission. Optimization of the device structure by incorporating TbPe as a fluorescent dopant in the 1,2-ADN layer further enhanced the performance. This optimized device exhibited a peak wavelength at 462 nm (2.68 eV), an ultralow turn-on voltage of 1.47 V (reaching 100 cd/m² at 1.97 V), and a maximum EQE of 3.25%. The operational lifetime was also significantly improved compared to a typical blue phosphorescent OLED, suggesting that the low energy of T1 plays a key role in enhanced stability. Moreover, even without the LiF electron injection layer, the device functioned, suggesting its robustness and the potential for further improved operational lifetime.

The results demonstrate the successful implementation of an upconversion mechanism in a blue OLED, using a common fluorescent emitter. The key to this achievement is the formation of a CT state at the D/A interface, facilitated by strong intermolecular interaction between the chosen donor (1,2-ADN) and acceptor (NDI-HF) materials. The efficient energy transfer from the CT state to the low-energy T1 of 1,2-ADN, followed by TTA to generate singlet excitons for emission, enabled a dramatic reduction in the turn-on voltage. The use of a fluorescent dopant further enhanced the EQE and efficiency. This work addresses the crucial challenge of creating highly efficient blue OLEDs with low operating voltages using readily available materials. The significantly improved operational lifetime compared to conventional phosphorescent OLEDs suggests that utilizing low-energy triplet states offers considerable advantages in terms of device stability.

This research successfully demonstrated an ultralow turn-on voltage (1.47 V) for blue light emission in an OLED, reaching a luminance of 100 cd/m² at 1.97 V, using a commonly used fluorescent emitter. The approach leverages a unique upconversion mechanism involving CT state formation and TTA. The findings significantly advance the development of low-voltage blue OLEDs and provide insights into controlling excitonic processes at organic semiconductor interfaces, which could have implications beyond OLEDs to areas like organic photovoltaics. Further research could focus on exploring different combinations of donor and acceptor materials, improving energy transfer efficiency, and enhancing the overall device stability.

While the achieved ultralow turn-on voltage is a significant advancement, the EQE of 3.25% is still relatively low compared to some high-efficiency blue OLEDs employing phosphorescent or TADF materials. However, these materials often suffer from stability issues that hinder their commercialization. Efficiency roll-off was observed at higher luminances, indicating that optimizing the device architecture to mitigate triplet-charge annihilation is a crucial area for future improvements. The choice of dopant material also influences the operational characteristics, necessitating further exploration of suitable dopants to fully optimize device performance for display applications. The simple D/A heterojunction structure used in this study, though stable compared to other types of devices, might limit performance compared to more complex multilayered devices.