Stable Thermally Activated Delayed Fluorescence‐Sensitized Red Fluorescent Devices through Physical Suppression of Dexter Energy Transfer

To date, thermally activated delayed fluorescence‐sensitized fluorescent organic light‐emitting diodes (TSF‐OLEDs) have undergone substantial research to achieve high efficiency and good operational stability in wide color gamut regions. Usually, to achieve a highly efficient TSF device, the Förster resonance energy transfer rate (kFRET) should be enhanced, whereas the Dexter energy transfer rate (kDET) should be suppressed. Even though highly efficient devices are achieved in all RGB color regions by satisfying the BT2020 requirements, achieving long device lifetimes is still challenging. Herein, a highly stable red‐TSF device is reported by adopting a new Dexter energy transfer suppressive layer (DSL) adjacent to the main emissive layer. Here, the DSL can improve the distance between the excitons generated from the host‐TADF layer and the final dopant (FD) of the TSF device, which allows for suppressing the kDET. Furthermore, the detailed device mechanistic pathways are analyzed by varying the DSL doping concentration with different thicknesses in different positions. Among the fabricated devices, the DSL‐TSF device manifested a longer operational lifetime (LT95) over 370 h at 5000 cd m‐2 and reduced efficiency roll‐off compared with TSF devices. Such long lifetime and high stability in DSL‐TSF OLEDs are owing to the decreased kDET than TSF devices.


Introduction
In recent days, organic light-emitting diodes (OLEDs) are evolving as promising optoelectronic devices owing to their high color purity, light weight, flexibility, wide viewing angle, and low cost. As of now, phosphorescent materials are used commercially in the OLED industries; however, they need highly cost material degradation, impacting the device's operational lifetime. [22,23] Hence, to achieve a highly efficient and stable TSF-OLED device with a longer lifetime, DET should be mostly minimized while FRET should be enhanced simultaneously.
On account of this issue, some approaches have been developed for reducing energy loss through the DET process, whereas the FRET process is unchanged. One approach is increasing intramolecular distance by inserting chemically inert-bulky groups into the fluorescent/TADF emitter. [24][25][26][27][28][29] For example, Kwon et al. reported a pure red TSF-OLED using the tert-butyl-substituted fluorescence emitter (4tBuMB) along with 4CzIPN and 4CzTPN TADF sensitizers. Such bulky group substitution increases the intramolecular distance and large spectral overlap with the sensitizers, which leads to a high FRET rate and reduced ISC/RISC cycles. As a result, the 4CzTPNbased OLED device has shown better efficiency and lifetime (LT 90 ) of 954 h at 3000 cd m -2 . [30] However, the DET process is not minimized well due to the low RISC rate and long-delayed fluorescence lifetime of 4CzTPN. To overcome this issue, Lee et al. recently reported a TADF molecule (12BTCzTPN) by modifying the 4CzTPN with a sulfur atom in two of the donors. The heavy atom effect of sulfur increases the RISC rate about 4.5 times that of 4CzTPN and suppresses the DET process. Which results in improved device efficiency and lifetime by 1.4-folds. [31] Apart from these chemical approaches, the improvement in the device architectures was also studied to alleviate the DET process. Liu et al. reported red TSF-OLEDs with an assistant emission layer (AEL) showing a long operational device lifetime (LT 95 ) of 900 h were achieved with corresponding CIE coordinates of (0.66, 0.33). [32] However, there is still a need for more improvements in highly efficient, stable red devices with long operational lifetimes for real-time applications.
In this work, we present a detailed mechanistic study of a TSF device with modified device architecture by inserting an additional layer called the 'DET suppress layer (DSL)' next to the final dopant without changing the emissive layer (EML) total thickness. The DSL consists of a host and a TADF sensitizer with a total thickness of 5 and 7 nm. Insertion of such DSL physically increases the distance between the excitons from the host-TADF system to the final emitter. This increased exciton passage distance fairly suppresses the short-range DET process while maintaining the FRET process. Since the DET is a distance-sensitive process, the thickness of DSL was restricted to 5 and 7 nm that are not too thin to fabricate. Additionally, by taking into account excitons distribution, the position of the DSL was adjusted between the TSF-emission layer (EML) and hole-blocking layer (HBL) rather than between the electronblocking layer (EBL) and TSF-EML. As a result, the optimized DSL-TSF devices have shown comparatively higher efficiency, 17%, and a longer operational lifetime of over 370 h at an initial luminescence of 5000 cd m -2 than the normal TSF devices.
The EL characteristics of TADF and TSF devices are shown in Figure 1, and the data are summarized in Table S1 (Supporting Information). In contrast, from Figure 1b, both TADF devices with 10% and 15% of 12BTCzTPN showed relatively low turn-on voltages of 2.7 V, indicating the low-energy barriers in all layers of the device. As observed in Figure 1c, the EL spectra of 10% and 15% doped devices were at 571 and 574 nm with FWHM of 94 and 95 nm, respectively. An increase in the concentration resulted in little red-shifted emission. These two TADF devices have shown maximum external quantum efficiency (EQE max ) of 8.3% and 6.9%, along with the operational lifetime of 206 and 187 h at an initial luminescence of 5000 cd m -2 (Figure 1d,e). However, among the two devices, the 10% doped TADF device shows better efficiency. In general, the higher concentration of TADF sensitizer causes more aggregation quenching effect, resulting in poor efficiency. [35] Accordingly, the TSF device was constructed using 10% of 12BTCzTPN and 0.7% of 4tBuMB and achieved a high EQE of 16.3% at an EL peak of 618 nm with the FWHM of 44 nm. Which showed a slightly shorter lifetime (LT 95 ) of 182 h at an initial luminescence of 5000 cd m -2 compared with those TADF devices (Figure 1c-e).
As aforementioned, we fabricated a TSF device by inserting an additional DSL consisting of 12BTCzTPN and DIC-TRZ. In order to attain better results, the position of DSL is important. It was fixed after analyzing the exciton distribution and the carrier recombination zone by evaluating a hole-only device (HOD) and electron-only device (EOD) of 12BTCzTPN. The results are shown in Figure 2. It is visible from Figure 2a that the J-V curve of the HOD is lower than that of the EOD, which indicates that the exciton distribution and recombination zone were mostly on the HBL side. Furthermore, sensing layer experiments were carried out to confirm the exciton distribution. [36,37] A red fluorescent material, 5,10,15,20-tetraphenylbisbenz [5,6]indeno[1,2,3cd:1′,2′,3′-lm]perylene (DBP) (the structure in Figure 2c) was employed as the sensing layer at three positions in EML such as EBL side, middle, and HBL side with the thickness of 0.6 nm as www.advmatinterfaces.de shown in Figure 2b. [38,39] The device was fabricated with the same total thickness as the TSF device and the device structure is: ITO (50 nm)/HATCN (HIL, 7 nm)/PCBBiF (HTL, 70 nm)/PCzAC (EBL, 10 nm)/DIC-TRZ: 10% 12BTCzTPN/DIC-TRZ: 10% 12BTC-zTPN (EML, x nm)/DBP (sensing layer, 0.6 nm)/DIC-TRZ: 10% 12BTCzTPN (EML, 29.4-x nm)/DDBFT (HBL, 5 nm)/BPPB: Liq (ETL, 60 nm)/LiF (EIL, 1.5 nm)/Al (100 nm). From Figure 2c, it is understood that the EL spectra were almost similar when the sensing layer (DBP) was moved from the EBL side to the middle side. A significant shift in EL spectra was observed when the sensing layer (DBP) was near HBL that further confirms that the exciton distribution was mainly on the HBL side. Based on this finding, we projected that the distribution of excitons in the TSF-DSL device would be on the DSL side as shown in Figure 2d.
The charge distribution results implemented the insertion of the additional layer on the HBL side fairly affect the device efficiency, and lifetime and also alters the emission spectra. Better outcomes would be obtained by inserting the DSL on the EBL side. TSF devices with DSL insertion on both the HBL and the EBL were fabricated under optimized device parameters and evaluated for comparable results. Initially, the DSL consisting of DIC-TRZ: 10% 12BTCzTPN with the thickness of 5 and 7 nm was inserted in between TSF-EML and HBL. To maintain the total EML thickness at 30 nm, the TSF-EML was fabricated with thicknesses of 25 and 23 nm along with 5 and 7 nm of DSL, respectively. The structure and the mechanism of the DSL-TSF-EML-based device have shown in Figure 3a,b. The evaluated results of these devices with 5 and 7 nm of DSL thickness have

www.advmatinterfaces.de
shown low turn-on voltage of 2.7 V and EQE max of 16.0% and 12.8% (Figure 3c,d). The two devices have almost similar EL peaks at 617 and 616 nm with FWHM of 48 and 54 nm, respectively. However, as shown in Figure 3e, the insertion of DSL on the HBL side results in poor color purity with intensified EL shoulder peak of 560-570 nm compared with the fabricated TSF. It was noticed that the shoulder peak is much improved in the device with the increased thickness of DSL. The respective operational lifetimes of devices are (LT 95 ) 213 and 255 h at initial luminescence of 5000 cd m -2 (Figure 3f). However, efficiency roll-off and device lifetime improved by 3.42 and 1.40 times compared with the TSF device, respectively. The summary of EL performances of TSF-DSL-HBL devices is tabulated in Table S2 (Supporting Information).
Subsequently, we fabricated and evaluated the device by repositioning the DSL between EBL and TSF-EML to minimize the emission of TADF sensitizer. Figure 4a shows the detailed device structure and the related mechanism was depicted in Figure 4b. All the devices have low turn-on voltage as shown in Figure 4c, which implies the low charge barriers in the device layers. The EBL-DSL-TSF-based devices with DSL thicknesses of 5 and 7 nm have achieved almost similar EQE max of 16.7% and 17.0%, respectively ( Figure 4d). As observed from Figure 4e, EL peak is not much affected compared with TSF and is found at 617 and 618 nm with FWHM of 45 and 45 nm, respectively. Moreover, the DSL with 7 nm thickness has shown a noticeably longer operational lifetime (LT 95 ) of 370 h than the 5 nm DSL-device with a lifetime of 280 h at an initial luminance of 5000 cd m -2 (Figure 4f). The most noteworthy lifetime improvement of the device was found to be 1.45 (7 nm DSL) times higher compared with the DSL on HBL side devices and 2.03 times higher than the TSF device. The EL characteristics are tabulated in Table 1.
Furthermore, we calculated the FRET radius (R 0 ) of 12BTC-zTPN using equation (1). R 0 signifies the distance at which the energy transition efficiency between the two molecules is 50%. [40] R n J 9000 128 where κ 2 is the orientation factor, Φ is the quantum yield of the fluorescence, N A is the Avogadro number, n is the refractive www.advmatinterfaces.de index, and J F is the overlap integral between donor emission and acceptor absorption. The calculated R 0 of 12BTCzTPN is 4.33 nm, which is lesser than the thickness of DSL (7 nm). Such a long distance between the two molecules than R 0 results in less FRET process. Hence, thicker DSL can also emit light with improved device lifetime and efficiency roll-off characteristics.
We have investigated the energy transfer processes relative to the excitons generated in TADF-sensitizer 12BTCzTPN as a donor and to the FD 4tBuMB as an acceptor in both TSF and DSL-TSF devices using time-resolved photoluminescence (TRPL) and PLQY measurements. [41] For the accurate k FRET and k DET evaluation, TSF thin film and DSL-TSF thin film were fabricated with the same doping concentration of both TADF sensitizer and FD. The total thickness was maintained at 30 nm, like the EML thickness of the device (Figure 5a). Where the chosen DSL thickness is 7 nm. The transient PL decay curves of prompt and delayed fluorescence of both TSF and DSL-TSF are shown in Figure 5b.
Both the TSF and DSL-TSF films displayed clear exponential decay curves exhibiting the nanoscale of prompt and the microscale of delayed fluorescence, as shown in Figure 5b. The energy-transfer rate constants k ISC , k RISC , k r, , and k nr of the TADF sensitizer are calculated using the equations S1-S6 (Supporting Information), and the corresponding results are tabulated in Table 2. Among these two films, the DSL-TSF film shows an improved delayed lifetime showing a more population of triplet excitons in TADF. Such an increase in triplet states indicates that the DSL insertion improves the delayed components in the TADF by suppressing triplet exciton transfer to the final emitter. The exciton dynamics in the DSL-TSF device compared with the normal TSF device shows increased ISC and RISC rates driven by the triplet population in the TADF. This generates more singlet excitons and enhances the FRET pro-cess to the final dopant. The improved radiative decay rate (k r ) of the TADF confirms the rapid transfer of the singlet excitons ( Table 2). Such enhanced efficiency by more generated singlet excitons is attributed to the improvement of device lifetime in the DSL-TSF. Generally, the TADF triplet excitons lifetime is in microseconds and milliseconds in the final emitter. Such longlived triplet excitons generated in the final emitter increase the probability of material degradation through TTA and TPA processes. But, the TADF sensitizer possesses a 10 3 -fold lowered triplet exciton lifetime than the final emitter, and the probability of the exciton annihilation processes is reduced, enhancing device operational stability. Additionally, the TADF triplet excitons are rapidly transferred to the final emitter by faster energy transfer, which reduces the duration of triplet excitons in the TADF sensitizer. Hence, we believe that the energy stored in the TADF triplets will not cause a serious impact on the device's lifetime. Such enhanced charge dynamic properties reduce the electrically induced device degradation, [19] resulting in reduced efficiency roll-off properties and an improved operational lifetime of 370 h (Table 1).
where k P and k D are prompt and delayed fluorescence rates, respectively. k r, S and k nr, T express the radiative rate constant of the singlet and the nonradiative rate constant of the triplet, respectively. k ISC and k RISC are intersystem crossingand reverse intersystem crossing rate constant, respectively.

www.advmatinterfaces.de
From Table 2, the DSL thin film showed slightly improved k FRET values of 1.16 × 10 8 s -1 and lowered k DET of 5.2 × 10 5 s -1 compared with TSF thin film. As per our ideology, DSL insertion could suppress DET with a maintained FRET process. A remarkable DET suppression of 2.62 times that of TSF thin film is observed. Such a low DET minimizes the triplet exciton transfer and notably reduces the long-lived triplet excitons in the final emitter and prohibits the destructive TTA and TPA processes. Hence, the final emitter degradation has significantly decreased, improving device stability (Table 1). This also induces better exciton generation that is transferred to the final emitter, resulting in improved device efficiency of 17%.
On the other hand, we have also evaluated the singlet and triplet excitons decay rate for the comparison of the exciton distribution in the TADF sensitizer by using the following equations (4) and (5). [45,46] dN t dt t t S FRET r n r,s where N S and N T are singlet and triplet exciton densities, respectively and k nr, S is the nonradiative decay rate constant of singlet excitons. The calculated values were plotted and shown in Figure 6a,b. As shown in Figure 6a, the singlet density of DSL-TSF and TSF devices did not significantly vary, which confirms the almost same FRET process that occurred in both devices. As a result, there was little variation in the singlet exciton density with respect to time. As can be seen from Figure 6b, a long-lived triplet exciton was detected in the DSL-TSF device when compared with the decay curves of the triplet excitons of both devices. It occurs because the DET process transfers fewer triplet excitons to the final emitter. The observation strengthens our findings that inserting a DSL enhances the device's lifetime by lowering the DET process and the efficiency roll-off characteristics (Figure 6c).

Conclusion
In summary, we have reported a clear mechanistic study of a TSF device that can remarkably suppress DET while maintaining the FRET process. This was achieved through the insertion of a "DSL" consisting of a host and a TADF sensitizer next to EML. The position of the DSL was optimized TSF-EML and EBL using exciton distribution evaluation. Such DSL insertion resulted in the noticeable suppression of DET by 2.62-fold compared with a standard TSF device with enhanced EQE max of 17.0%. The DSL-TSF showed a pure red emission at 618 nm with an FWHM of 45 nm, and the corresponding CIE coordinates are (0.63, 0.37). Remarkably, a more extended device operational lifetime (LT 95 ) of 370 h at an initial luminance of 5000 cd m -2 was achieved. We anticipate that the design approach described in this work will serve as an example for more OLED devices with longer lifetimes.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.  Figure 6. The decay rate of: a) singlet-, b) triplet exciton density depending on time, c) EQE-current density property of SDL (7 nm)/TSF-EML device and TSF device.