Polymeric Hole Transport Materials for Red CsPbI3 Perovskite Quantum-Dot Light-Emitting Diodes

In this study, the performances of red CsPbI3-based all-inorganic perovskite quantum-dot light-emitting diodes (IPQLEDs) employing polymeric crystalline Poly(3-hexylthiophene-2,5-diyl) (P3HT), poly(9-vinycarbazole) (PVK), Poly(N,N′-bis-4-butylphenyl-N,N′-bisphenyl)benzidine (Poly-TPD) and 9,9-Bis[4-[(4-ethenylphenyl)methoxy]phenyl]-N2,N7-di-1-naphthalenyl-N2,N7-diphenyl-9H-fluorene-2,7-diamine (VB-FNPD) as the hole transporting layers (HTLs) have been demonstrated. The purpose of this work is an attempt to promote the development of device structures and hole transporting materials for the CsPbI3-based IPQLEDs via a comparative study of different HTLs. A full-coverage quantum dot (QD) film without the aggregation can be obtained by coating it with VB-FNPD, and thus, the best external quantum efficiency (EQE) of 7.28% was achieved in the VB-FNPD device. We also reported a standing method to further improve the degree of VB-FNPD polymerization, resulting in the improved device performance, with the EQE of 8.64%.

To date, Poly-TPD has been selected as the hole transporting material in most of IPQLED studies [12][13][14][15][16][17][18][19][20][21]. Poly(triaryl)amine (PTAA) also shows a high hole mobility, which makes it another good choice for the hole transporting materials. The defect during the formation process of the QD films could be reduced by PTAA [9,10], which is effective in enhancing the radiative recombination. The more crystalline Poly(3-hexylthiophene-2,5-diyl) (P3HT) has a relatively higher hole mobility than that of noncrystalline organic HTLs and is often used as light-harvesting and hole transporting materials for CsPbI 3 -based solar cells [25]. Our previous report demonstrated that a thermal crosslinkable HTL, 9,methoxy]phenyl]-N2,N7-di-1-naphthalenyl-N2,N7diphenyl-9H-fluorene-2,7-diamine (VB-FNPD), also provides excellent hole mobility and improves the interface between the HTL and the CsPbBr 3 QD film [11]. On the other hand, the studies of the HTLs for deep red CsPbI 3 -based IPQLEDs are still lacking. Table 1 shows only Poly-TPD and PTAA have been used as the HTLs in the CsPbI 3 -based IPQLEDs. No literature reported the CsPbI 3 -based IPQLEDs using VB-FNPD and P3HT as the HTLs. Therefore, a comparative study of different HTLs for the CsPbI 3 -based IPQLEDs is necessary. Herein, we studied the performance of CsPbI 3 -based IPQLEDs employing P3HT, PVK, Poly-TPD and VB-FNPD as the HTLs. Meanwhile, a dense and smooth CsPbI 3 QDs film can be achieved using VB-FNPD HTLs, which are an important factor for the device performance of the IPQLED. We then demonstrated highly bright and efficient CsPbI 3 IPQLED based on VB-FNPD HTLs, achieving an external quantum efficiency (EQE) of 8.64%. Therefore, we believe that our results may promote the development of device structures and hole transporting materials to achieve stable and low-cost IPQLEDs.

Synthesis of CsPbI 3 QDs
Cs 2 CO 3 (200 mg) was loaded into a 25 mL three-neck flask, along with ODE (9 mL) and OA (0.75 mL), and then stirred and degassed at 120 • C for 30 min under nitrogen flow to obtain a transparent Cs-oleate precursor. The Pb precursor solution was prepared by dissolving 0.09 M of PbI 2 in 30 mL ODE, 3 mL of OA and 3 mL OAM and then stirring and degassing at 120 • C under nitrogen flow. After PbI 2 was all dissolved, the temperature was increased to 150 • C, and then a 0.8 mL Cs-oleate precursor was quickly injected into the Pb precursor solution. After 5 s, the reaction was cooled on an ice bath, and red CsPbI 3 QD crude was obtained. Then n-octylammonium iodide (0.2 mmol) dissolved in toluene (4 mL), as a capping agent was added into the crude. Subsequently, as-prepared crude solution and methyl acetate (16 mL) were centrifugated at 12,000 rpm for 15 min. The precipitate was collected and loaded in 8 mL of hexane and methyl acetate (1:3 v/v), and the solution was centrifuged at 12,000 rpm for 10 min. The precipitate was collected and dispersed in octane (2 mL) and centrifuged for 5 min at 12,000 rpm. Finally, the supernatant was collected and stored at 4 • C.

Device Fabrication
The IPQLEDs were constructed with the architecture of indium tin oxide (ITO)/ PEDOT:PSS (40 nm)/ HTLs (~50 nm)/ CsPbI 3 QD (~40 nm)/ TPBi (40 nm)/ LiF (1 nm)/Al (100 nm). Here, P3HT, PVK, Poly-TPD and VB-FNPD were used as the HTLs. The patterned ITO substrates were wet-cleaned and then O 2 Plasma-cleaned. After cleaning, PEDOT:PSS was spin-coated at 8000 rpm for 40 s on the substrate and annealed at 130 • C for 15 min. Then, the samples were loaded to N 2 -filled glove box to deposit HTLs and CsPbI 3 QDs. All HTLs were spin-coated with a concentration of 4 mg/mL on PEDOT:PSS and then heated at 100 • C for 5 min. The thickness of each HTL was controlled at~50 nm by adjusting the spinning speed. Before heating, VB-FNPD was held standing still for 0, 20, 40 and 60 min and then heated at 100 • C for 5 min and annealed at 170 • C for 30 min for thermal crosslinking. The CsPbI 3 QDs were spin-coated with a concentration of 40 mg/mL at 2000 rpm for 60 s. TPBi, LiF and Al cathode were deposited by a thermal evaporation using a shadow mask to define the device area of 2 × 2 mm 2 .

Characterization
Electroluminescence and impedance characteristics were measured through computercontrolled LQ-100R spectrometer (Enlitech, Kaohsiung, Taiwan) and Material Lab XM (SOLARTRON analytical, Leicester, UK), respectively. The absorbance and photoluminescence (PL)/photoluminescence quantum yield (PLQY) were measured using UV-visible spectrophotometer (V-770, JASCO, Tokyo, Japan) in Table 2 and fluorescence spectrophotometer (F-7000, Hitachi, Tokyo, Japan), respectively. The surface roughness was measured using an atomic force microscope (AFM, Bruker, Billerica, MA, USA). The electron microscopy images were obtained by HRTEM (JEM-2100, JEOL, Tokyo, Japan) and FESEM (JSM-7610F, JEOL, Tokyo Japan), respectively.  Figure 1a shows the planar SEM image of the CsPbI 3 QD film spun on the glass substrate. Highly dense surface and good crystalline of the CsPbI 3 QD film can be obtained Polymers 2021, 13, 896 4 of 10 without obvious aggregations. Such morphology may be attributed to the well-dispersed and high-stability suspensions in the as-synthesized QD dispersions, as shown in the insert in Figure 1. The PL spectrum (Figure 1b) of the CsPbI 3 QDs film shows a brightly red luminescence at 682 nm with a narrow Full width at half maximum (FWHM) of 35 nm, implying a high color purity and preferred optical property. The absorption edge in the absorption spectrum is close to its emission peak, which agrees with previous reports [22][23][24][25]. TEM image shows as-synthesized CsPbI 3 QDs are cubic shaped and welldispersed in octan, with an average size of 10.8 nm (Figure 1c,d). All abovementioned characterization techniques evidently exhibit that the CsPbI 3 QD dispersion solutions and QD solid films with uniform size and distribution have been successfully obtained.  Figure 1a shows the planar SEM image of the CsPbI3 QD film spun on the glass substrate. Highly dense surface and good crystalline of the CsPbI3 QD film can be obtained without obvious aggregations. Such morphology may be attributed to the well-dispersed and high-stability suspensions in the as-synthesized QD dispersions, as shown in the insert in Figure 1. The PL spectrum (Figure 1b) of the CsPbI3 QDs film shows a brightly red luminescence at 682 nm with a narrow Full width at half maximum (FWHM) of 35 nm, implying a high color purity and preferred optical property. The absorption edge in the absorption spectrum is close to its emission peak, which agrees with previous reports [22][23][24][25]. TEM image shows as-synthesized CsPbI3 QDs are cubic shaped and well-dispersed in octan, with an average size of 10.8 nm (Figure 1c,d). All abovementioned characterization techniques evidently exhibit that the CsPbI3 QD dispersion solutions and QD solid films with uniform size and distribution have been successfully obtained.   [12,25,29]. Figure 2b-d show the device performance of the CsPbI3 IPQLED using different HTLs. LUMO levels of all HTLs are much higher than those of the QD layer, resulting in good electron blocking ability in all HTLs (Figure 2a). HOMO levels of all HTLs are higher than those of the QD layer, indicating that reducing hole injection barrier is preferred to the HTL with the lower HOMO level. Therefore, the tendencies of the current in devices' different HTLs correspond with the HOMO level of their HTL (Figure 2c). The PVK device shows the highest current, because of the lowest HOMO in PVK, which is agreed with the lowest impedance  The highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) levels of all layers can be referred to the results in [12,25,29]. Figure 2b-d show the device performance of the CsPbI 3 IPQLED using different HTLs. LUMO levels of all HTLs are much higher than those of the QD layer, resulting in good electron blocking ability in all HTLs (Figure 2a). HOMO levels of all HTLs are higher than those of the QD layer, indicating that reducing hole injection barrier is preferred to the HTL with the lower HOMO level. Therefore, the tendencies of the current in devices' different HTLs correspond with the HOMO level of their HTL (Figure 2c). The PVK device shows the highest current, because of the lowest HOMO in PVK, which is agreed with the lowest impedance ( Figure S1). In contrast, the lowest current in the P3HT device is caused by the highest HOMO level and hole injection barrier, leading to the highest impedance ( Figure S1) and turn-on voltage (biased voltage at 1 cd/m 2 ), as shown in Figure 2b. Similar HOMO levels in VB-FNPD and Poly-TPD lead to their same turn-on voltages, but the excellent radiative recombination efficiency in the VB-FNPD device gives it higher EQE. In addition, the PVK device has the highest current, but it simultaneously shows the lowest EQE (Figure 2d), which may be caused by inefficient radiative recombination in the QD layers, leading to its higher turn-on voltage than that of VB-FNPD and Poly-TPD devices ( Figure  2b). It is interesting to know what dominates as the carrier recombination efficiency for each device.

Results and Discussion
Polymers 2021, 13, x FOR PEER REVIEW 5 of 10 ( Figure S1). In contrast, the lowest current in the P3HT device is caused by the highest HOMO level and hole injection barrier, leading to the highest impedance ( Figure S1) and turn-on voltage (biased voltage at 1 cd/m 2 ), as shown in Figure 2b. Similar HOMO levels in VB-FNPD and Poly-TPD lead to their same turn-on voltages, but the excellent radiative recombination efficiency in the VB-FNPD device gives it higher EQE. In addition, the PVK device has the highest current, but it simultaneously shows the lowest EQE (Figure 2d), which may be caused by inefficient radiative recombination in the QD layers, leading to its higher turn-on voltage than that of VB-FNPD and Poly-TPD devices (Figure 2b). It is interesting to know what dominates as the carrier recombination efficiency for each device. In fact, the determination of the carrier recombination efficiency can be easily observed by the naked eye. Figure 3a shows the photograph of CsPbI3 QDs films spun on each HTL, in which the VB-FNPD film shows brighter than others. The thicknesses of all QDs films were around 40 nm, measured by Alpha-Step. Hence, the brightness of the QD solids should be attributed to the degree of the aggregation on the different HTL surfaces, rather than the film thickness. When the well-organized array of the colloidal CsPbI3 QDs is formed on the surface of the VB-FNPD films without the QD aggregations, the lightinduced exciton is limited in a QD nanoparticle to increase the quantum confinement effect, resulting in the improved radiative recombination, as illustrated in Figure 3b. In contrast, the light-induced exciton can transport between nanoparticles, due to the QD aggregations, leading to the increased dissociation possibility of the exciton prior to its radiative decay [30,31]. It is the reason why the brightness of VB-FNPD film is much stronger than that of other films, which is in good agreement with the device results (Figure 2d). The summary of PLQYs for CsPbI3 QDs layers spun on different HTLs are listed in Table 1. The PLQY of CsPbI3 QDs layer on the glass is higher than the PLQY of those spun on each In fact, the determination of the carrier recombination efficiency can be easily observed by the naked eye. Figure 3a shows the photograph of CsPbI 3 QDs films spun on each HTL, in which the VB-FNPD film shows brighter than others. The thicknesses of all QDs films were around 40 nm, measured by Alpha-Step. Hence, the brightness of the QD solids should be attributed to the degree of the aggregation on the different HTL surfaces, rather than the film thickness. When the well-organized array of the colloidal CsPbI 3 QDs is formed on the surface of the VB-FNPD films without the QD aggregations, the light-induced exciton is limited in a QD nanoparticle to increase the quantum confinement effect, resulting in the improved radiative recombination, as illustrated in Figure 3b. In contrast, the light-induced exciton can transport between nanoparticles, due to the QD aggregations, leading to the increased dissociation possibility of the exciton prior to its radiative decay [30,31]. It is the reason why the brightness of VB-FNPD film is much stronger than that of other films, which is in good agreement with the device results (Figure 2d). The summary of PLQYs for CsPbI 3 QDs layers spun on different HTLs are listed in Table 1. The PLQY of CsPbI 3 QDs layer on the glass is higher than the PLQY of those spun on each HTL, which may be because the exciton dissociation is suppressed at the insulated glass [32]. On the other hand, the full-coverage QD films confirm the carrier combination. The films with the QD aggregations provide a leakage path, which is the reason that the current in the Poly-TPD device is higher than that in the VB-FNPD device (Figure 2c). The best performance of 7.28% was achieved in the VB-FNPD device.
Polymers 2021, 13, x FOR PEER REVIEW 6 of 10 HTL, which may be because the exciton dissociation is suppressed at the insulated glass [32]. On the other hand, the full-coverage QD films confirm the carrier combination. The films with the QD aggregations provide a leakage path, which is the reason that the current in the Poly-TPD device is higher than that in the VB-FNPD device (Figure 2c). The best performance of 7.28% was achieved in the VB-FNPD device. To further improve the device performance, the different standing times were introduced into the VB-FNPD film preparation. Figure 4 shows current-voltage-luminance characteristics, EQE and the normalized electroluminescence (EL) spectrum of the devices prepared by different standing times. The performances of all devices with the standing treatment show better than that of the device without the treatment. EL spectrum shows an emission peak at 680 nm with a narrow FWHM of 32 nm, indicating high color purity. The EL peak position is close to the PL spectrum, which can be attributed to carrier recombination in the QD films. Figure 5 shows the AFM images of the VB-FNPD films with the different standing times. The AFM phase image exhibits that light and dark colors are alternately and uniformly distributed on the VB-FNPD film surface without the standing treatment, indicating two-phase coexistence [33] and low degree of polymerization. With the increase in the standing times, the deepened colors and the larger domain sizes on the phase images can be found, which could be attributed to the increased degree of the polymerization. Therefore, the reduced surface roughness can be seen in the AFM topography images, leading to the improved hole transporting characteristic and the device performance. Thus, the highest EQE of 8.64% in the VB-FNPD devices treated for 60 min were achieved. To further improve the device performance, the different standing times were introduced into the VB-FNPD film preparation. Figure 4 shows current-voltage-luminance characteristics, EQE and the normalized electroluminescence (EL) spectrum of the devices prepared by different standing times. The performances of all devices with the standing treatment show better than that of the device without the treatment. EL spectrum shows an emission peak at 680 nm with a narrow FWHM of 32 nm, indicating high color purity. The EL peak position is close to the PL spectrum, which can be attributed to carrier recombination in the QD films. Figure 5 shows the AFM images of the VB-FNPD films with the different standing times. The AFM phase image exhibits that light and dark colors are alternately and uniformly distributed on the VB-FNPD film surface without the standing treatment, indicating two-phase coexistence [33] and low degree of polymerization. With the increase in the standing times, the deepened colors and the larger domain sizes on the phase images can be found, which could be attributed to the increased degree of the polymerization. Therefore, the reduced surface roughness can be seen in the AFM topography images, leading to the improved hole transporting characteristic and the device performance. Thus, the highest EQE of 8.64% in the VB-FNPD devices treated for 60 min were achieved.

Conclusions
In conclusion, polymeric hole transport materials employed for red CsPbI3 IPQLEDs have been demonstrated. The band-aligned and aggregation characteristics of the CsPbI3

Conclusions
In conclusion, polymeric hole transport materials employed for red CsPbI 3 IPQLEDs have been demonstrated. The band-aligned and aggregation characteristics of the CsPbI 3 layers deposited on P3HT, PVK, Poly-TPD and VB-FNPD HTLs were discussed. A fullcoverage QD film without the aggregation can be obtained on the VB-FNPD films, and thus, the best performance was 7.28% in the VB-FNPD device. One of the key issues associated with the utilization of thermal-crosslinking polymer thin films is the control of their alignment and orientation. A standing method of increasing the degree of VB-FNPD polymerization was also presented, resulting in the improved device performance with the EQE up to 8.64%.