Trap-Controlled White Electroluminescence From a Single Red-Emitting Thermally Activated Delayed Fluorescence Polymer

Single white-emitting polymers have been reported by incorporating the second-generation carbazole dendron into the side chain of a red-emitting thermally activated delayed fluorescence (TADF) polymer. Due to the prevented hole trap effect, in this case, excitons can be generated simultaneously on the polymeric host and the red TADF dopant to give a dual emission. Consequently, a bright white electroluminescence is achieved even at a dopant loading as high as 5 mol.%, revealing a maximum luminous efficiency of 16.1 cd/A (12.0 lm/W, 8.2%) and Commission Internationale de l'Eclairage (CIE) coordinates of (0.42, 0.32). The results clearly indicate that the delicate tuning of charge trap is a promising strategy to develop efficient single white-emitting polymers, whose low-band-gap chromophore content can be up to a centesimal level.


INTRODUCTION
Single white-emitting polymers (SWPs) have attracted much attention owing to their potential applications in flat-panel displays and solid-state lightings (Reineke et al., 2013). In this case, several fluorescence, phosphorescence, and/or thermally activated delayed fluorescence (TADF) chromophores with either two complementary colors (blue and yellow) or three primary colors (blue, green, and red) are covalently incorporated into a single polymeric host at the same time so as to generate white electroluminescence (EL) (Liu et al., 2007a,b;Shao et al., 2012;Li et al., 2017;Wang et al., 2017). Compared with the physical blend systems, the undesirable phase segregation can be avoided effectively, leading to improved device performance as well as good spectral stability (Tu et al., 2004). However, the molar ratio of the long-wavelength chromophores in SWPs is often required to be in the range of one ten thousandth to one thousandth (Liu et al., 2007a;Shao et al., 2012;Wang et al., 2017). The extremely low doping concentration is difficult to be controlled during polymerization, which may bring about batch-to-batch variation for the synthesis of SWPs, and thus poor device reliability and reproducibility.
We note that there are few studies on how to address such an issue, although a great progress has been made on the power efficiency of SWPs recently (Shao et al., 2018). Here, we report TADF-based SWPs, whose low-band-gap chromophore content can be raised up to a centesimal level. This is achieved by simply introducing the second-generation carbazole dendron into the side chain of a previously-reported red-emitting TADF polymer PCzDMPE-R5.0 (Figure 1) GRAPHICAL ABSTRACT | Single white-emitting polymers (SWPs) were achieved by simply introducing the second-generation carbazole dendron to suppress hole trap effect on the dopant. (Yang et al., 2019). Because of the suppressed hole trap effect on the dopant, an interesting dual emission originating from both host and dopant is observed under the electrical excitation for all the resultant SWPs (D2-PCzDMPE-R2.5 ∼ D2-PCzDMPE-R10). Among them, D2-PCzDMPE-R5.0 gives a more balanced white EL, revealing a maximum luminous efficiency of 16.1 cd/A (12.0 lm/W, 8.2%) and Commission Internationale de l'Eclairage (CIE) coordinates of (0.42, 0.32). The results clearly indicate that the delicate tuning of charge trap is a promising strategy to develop efficient SWPs with a high loading of longwavelength chromophores.

Synthesis and Characterization
The synthetic route of the TADF-based SWPs is depicted in Scheme 1. Starting from the second-generation oligocarbazole D2, two successive N-alkylated reactions were carried out to afford the key monomer M1. Combined with other two comonomers M2 and M3, then a Suzuki polymerization was adopted to produce the target polymers D2-PCzDMPE-R2.5, D2-PCzDMPE-R5.0, D2-PCzDMPE-R7.5, and D2-PCzDMPE-R10. Their number-average molecular weights and polydispersity indexes were determined to be 63-101 kDa and 1.58-1.66, respectively (Table 1). And the actual content of the red TADF dopant incorporated into polymer can be calculated using their 1 H NMR spectra. As one can see in Figure S1, the characteristic signals of δ8.46 and 8.28 are subjected to the anthraquinone segment in the red TADF emitter, while the peak at about δ2.31 is from the methyl group in the 3,3 ′ -dimethyldiphenyl ether building block. By comparing their relative integrals, the red dopant loading is estimated to be in the range of 2.4-10.0%, very close to the feed ratio (Table 1). This implies that the red TADF emitter has been successfully bonded into the SWPs during polymerization.
In addition, they all exhibit a decomposition temperature (T d : corresponding to a 5% weight loss) of 455-470 • C and a glass transition temperature (T g ) of 259-266 • C ( Figure S2), much higher than those of PCzDMPE-R5.0 (T d = 417 • C, T g = 94 • C) (Yang et al., 2019). The introduced oligocarbazole functionalized with tert-butyl groups may be responsible for the improved thermal stability of D2-PCzDMPE-R2.5 ∼ D2-PCzDMPE-R10 (Zhao et al., 2015. Also, it contributes to their good solubility in common organic solvents (toluene, chlorobenzene, and chloroform etc.), which ensures the generation of high quality films via spin coating.

Polymers
Red emitter content in the polymers (mol%) Feed ratio Actual content a

2015)
. And the charge transfer (CT) absorption related to the red TADF dopant seems to be weak but distinguishable, lying in the range of 400-550 nm. Due to the incomplete energy transfer, moreover, a dual emission is detected for all the polymers. One is from the blue polymeric host, the other is from the red TADF dopant, whose intensity is found to be gradually increased with the increasing doping concentration.
Unlike the red-emitting PCzDMPE-R5.0, we note, D2-PCzDMPE-R5.0 obtains a dual emission from both the polymeric host and small-molecular TADF dopant (Figure 4). Given the same feed ratio, the additional second-generation carbazole dendron plays an important role on the observed difference. As  for PCzDMPE-R5.0, there exists a strong hole trap effect owing to the much deeper highest occupied molecular orbital (HOMO) level of the polymeric host relative to the red TADF dopant (−5.92 eV vs. −5.20 eV). In this case, the injected holes cannot be stored on host but trapped by dopant completely, while electrons are injected into dopant via an electrostatic attraction (Adachi et al., 2000;Tessler et al., 2000;Lane et al., 2001;Gong et al., 2003). Then excitons are generated directly on dopant, and only red emission appears in the EL spectrum of PCzDMPE-R5.0 ( Figure 4A). By contrast, the incorporated oligocarbazole has led to a distinct HOMO upshift from −5.92 to −5.47 eV for the polymeric host in D2-PCzDMPE-R5.0 ( Figure S6). Benefitting from the suppressed hole trap, holes can be accumulated either on host or on dopant. After recombination with the injected electrons via an electrostatic attraction, two classes of excitons are able to be formed on both host and dopant, resulting in a dual emission and thus white EL ( Figure 4B) (Liu et al., 2005;Farmer et al., 2011;Li et al., 2014). To avoid the above-mentioned aggregation induced TTA in neat films, doped devices were further assembled with D2-PCzDMPE-R5.0 as an example. When it is doped into 1,3bis(9H-carbazol-9-yl)benzene (mCP) at a 30 wt.% concentration, the current efficiency, power efficiency and EQE are optimized to be 16.1 cd/A, 12.0 lm/W and 8.2%, respectively (Figure 5 and Figure S7). Meanwhile, the EL spectrum remains nearly unchanged, accompanied by similar CIE coordinates of (0.42, 0.32) to the non-doped device. Although the obtained performance is moderate, the loading of long-wavelength dopant here is as high as 5 mol.%, one or two order magnitude higher than those of previously-reported SWPs (Chuang et al., 2007;Liu et al., 2007a;Luo et al., 2007;Shao et al., 2012;Wang et al., 2017). This is very instructive when trying to solve the batch-to-batch variation in material synthesis.

CONCLUSIONS
In summary, a red to white conversion has been demonstrated by incorporating the second-generation carbazole dendron into the side chain of a red-emitting TADF polymer. Benefitting from the elevated HOMO level of the polymeric host, the hole trap effect between host and dopant is reasonably weakened. As a consequence, a dual emission from both host and dopant is observed simultaneously, leading to a bright white EL even at a 5 mol.% dopant content. This work provides an effective strategy to improve the loading of long-wavelength chromophores up to a centesimal level, which will shed light on the development of SWPs showing not only high power efficiency but also good reproducibility.

Measurements and Characterization
1 H NMR and 13 C NMR spectra were recorded with a Bruker Avance 400 spectrometer or Bruker Avance 500 spectrometer. MALDI/TOF (matrix-assisted laser desorption ionization/timeof flight) mass spectra were performed on an AXIMA CFR MS apparatus (COMPACT). Molecular weights of the polymers were determined by Gel permeation chromatography (GPC) in tetrahydrofuran (THF) using polystyrene as the standard. Thermal gravimetric analysis (TGA) and differential scanning calorimetry (DSC) were performed under a flow of nitrogen with PerkinElmer-TGA 7 and PerkinElmer-DSC7 systems, respectively. UV-vis absorption and PL spectra were measured with a PerkinElmer Lambda 35 UV-vis spectrometer and a PerkinElmer LS 50B spectrofluorometer, respectively. By dropping 0.5 ml solution of polymers dissolved in toluene to an optical colorimetric dish and then drying under vacuum, the films used for PLQY measurement are formed on the walls of the colorimetric dish. The film PLQYs were measured using a quantum yield measurement system (C10027, Hamamatsu Photonics) excited at 350 nm under argon protection. PLQYs are calculated from the area integral ratio of emission to absorption. And the transient PL spectra were carried out with Edinburgh fluorescence spectrometer (FLS980). The HOMO and lowest unoccupied molecular orbital (LUMO) levels were estimated from the cyclic voltammetry (CV), which was performed on a CHI660a electrochemical analyzer with Bu 4 NClO 4 (0.1 mol/L) as the electrolyte at a scan rate of 100 mV/s. A glass carbon electrode, a saturated calomel electrode, and a Pt wire were used as the working electrode, the reference electrode, and the counter electrode, respectively. All the potentials were calibrated by ferrocene/ferrocenium (Fc/Fc + ). HOMO = -e (Eox onset + 4.8 V), LUMO = HOMO + E g , where Eox onset is the onset value of the first oxidation wave and the E g is the optical bandgap estimated from the absorption onset.

Device Fabrication and Testing
The indium tin oxide (ITO) (20 per square) substrates were cleaned with acetone, detergent, distilled water and then in an ultrasonic solvent bath. After baking in a heating chamber at 130 • C for 2 h, the ITO-glass substrates were treated with UVozone for 25 min. Firstly, PEDOT:PSS (Batron-P4083, Bayer AG) was spin-coated on top of the ITO at a speed of 5,000 rpm for 60 s, and baked at 120 • C for 45 min. After transferred into a nitrogen-filled glove-box, subsequently, solutions of polymers in toluene were spin-coated on PEDOT:PSS as the EML at a speed of 1,500 rpm for 60 s, and annealed at 80 • C for 0.5 h. Finally, the other layers including TmPyPB (50 nm), LiF (1 nm) and Al (100 nm) were deposited in a vacuum chamber at a base pressure of >4 × 10 −4 Pa. The EL spectra and CIE coordinates were measured using a CS2000A spectra colorimeter. The current-voltage and brightness-voltage curves of devices were measured using a Keithley 2,400/2,000 source meter and a calibrated silicon photodiode. All the measurements were carried out at room temperature under ambient conditions.