Polyethylenimine-ethoxylated dual interfacial layers for highly efficient and all-solution-processed inverted quantum dot light-emitting diodes

1. Introduction

Colloidal quantum dots (QDs) have attracted considerable attention as emitting materials in next-generation display systems because of their high photoluminescence (PL) quantum yield, outstanding narrow-band and widely tunable emission, and low-cost solution-based processability [1–6]. Electrically driven quantum dot light-emitting diodes (QLEDs) have been intensively investigated as an alternative to organic light-emitting diodes (OLEDs) because of their high color purity and low-cost solution-based processability [7–9]. The efficiency of QLEDs has substantially improved over the past two decades. The performance of state-of-the-art QLEDs is comparable to that of commercial OLEDs in terms of theoretical external quantum efficiency (EQE) [10–13] because of the advancement in the synthesis of high-quality QD materials [14–16] and sophistication of device structures [17,18].

So far, QLEDs having conventional or inverted structures have been fabricated. Conventional QLEDs have been widely applied in various optoelectronic devices. However, inverted QLEDs are advantageous because its indium tin oxide (ITO) cathode with pixel patterns can be directly connected to an n-type thin film transistor backplane with a low driving voltage [18–20]. Moreover, zinc oxide (ZnO) nanoparticle films are typically used as electron transport layers (ETLs), and their mobility in inverted QLEDs can be improved by annealing at high temperature after coating the ETLs on the ITO substrate [18]. Inverted QLEDs are widely fabricated by vacuum processes, and solution-based processes are rarely conducted. However, solution-based processes are advantageous because they are cost effective for large-scale production, promote low material loss, are compatible with flexible substrates, and facilitate large-area processing [8,21].

However, the efficiency of all-solution-processed inverted QLEDs is lower than that of the conventional QLEDs. One of the reasons for the low efficiency is the destruction of the QD emitting layers (EMLs) by typical hole transport layer (HTL) solvents, such as toluene, chlorobenzene, and chloroform. Conventional QDs have hydrophobic aliphatic ligands (e.g., oleic acid) and are readily soluble in non-polar HTL solvents, leading to the mixing of the QD EMLs and the neighboring HTLs. The rough surface caused by EML dissolution is known to cause a decrease in the efficiency and stability of the devices because of non-uniform exciton recombination and leakage current in the EML [22–24]. Various approaches have been attempted to address the issues caused by solvent damage, including the use of orthogonal solvents [25–27], surface modification of the QD EML [24], and addition of a protective layer on the QD EML [23]. Another challenge is the charge imbalance caused by excess electrons in QLEDs owing to higher electron mobility through the ZnO nanoparticle ETL than the hole mobility through the common organic HTLs. Excess electrons cause exciton quenching by Auger recombination, thereby decreasing efficiency [28–30]. The injection and transport of electrons have been effectively suppressed by doping various elements (e.g., Al, Li, Mg, Ga) onto ZnO [31–34] or by introducing an insulating interfacial layer, such as poly [(9,9-bis(30-(N,N-dimethylamino)propyl)-2,7-fluorene)-alt-2,7-(9,9-ioctylfluorene)] (PFN) [35], poly(methyl methacrylate) (PMMA) [10], polyethylenimine (PEI) [23,36], and aluminum oxide [37]. However, few studies have been conducted on all-solution-processed inverted QLEDs.

Herein, QD EML was sandwiched between polyethylenimine-ethoxylated (PEIE) layers to overcome the solvent damage and charge imbalance in all-solution-processed inverted QLEDs. First, PEIE layers were formed on the QD EML as an EML protecting layer (EPL) to offer protection from the HTL solvents. The PEIE EPL thickness was optimized, and solvent resistance was evaluated. The integrity of the device was confirmed with and without the presence of PEIE layers. The surface hydrophilicity of the QD EML was also determined in the absence and presence of the PEIE EPL. Next, the PEIE EPL was introduced between the QD EML and the ZnO ETL as an electron-blocking layer (EBL) to suppress electron injection in devices. Finally, inverted QLEDs with different PEIE EBL thicknesses were fabricated via an all-solution process, and their optical performance was evaluated.

2. Experimental section

2.1 Synthesis of green-emitting CdSe@ZnS/ZnS QDs

Green-emitting CdSe@ZnS/ZnS QDs with a chemical composition gradient were prepared according to methods described in literature [16]. First, 0.14 mM of cadmium acetate, 3.14 mM of zinc oxide, and 7 mL of oleic acid were mixed in 15 mL of 1-octadecene and degassed at 110 °C under vacuum for 1 h to prepare cadmium oleate. The reactor temperature was then increased to 310 °C in an inert environment (nitrogen gas). Next, 1.5 mM Se and 2.5 mM S were dissolved in 3 mL TOP to prepare a stock solution that was injected into the reactor and stirred vigorously for 10 min. To synthesize alloyed CdSe@ZnS/ZnS QDs, a sulfur stock solution (1.6 mM of S dissolved in 2.4 mL of OED) was injected sequentially and the reaction was maintained at 270 °C for 10 min. The crude solution was precipitated with ethanol and centrifuged to remove excess free ligands, unreacted precursors, and impurities. Finally, the precipitate was re-dispersed in a mixture of hexane and octane for further analysis, including UV-vis, PL, TEM, and for device fabrication. The green-emitting QDs exhibited excellent PL quantum yield (QY) of 90% in solution. The PL spectra show the peak wavelength at 523 nm with a narrow full width at half-maximum (FWHM) of 20 nm. A distinguishable excitonic absorption peak is also observed and the QDs exhibited uniform size distribution of 13.5 nm on average, as shown in Fig. 1.

 

Fig. 1. (a) Absorption and PL spectra of green-emitting QD solution in hexane/octane. (inset) Image of QD solution under UV light at 365 nm. (b) TEM image of QDs.

Download Full Size | PPT Slide | PDF

2.2 QLEDs fabrication of inverted QLEDs

A patterned ITO glass substrate with a sheet resistance of ∼15 Ω sq⁻¹ was cleaned sequentially with deionized water (DIW), acetone, and methanol. The colloidal ZnO NP solution (200 mg/mL) in ethanol was spin-coated (1500 rpm, 60 s) on the ITO and annealed at 180 °C for 20 min. The PEIE concentration for EBL was varied from 0 to 0.5 wt% in 2-methoxyethanol, and the PEIE solution was spin-coated (3000 rpm, 30 s) onto the ZnO film, followed by drying at 150 ˚C for 5 min. Then, the green CdSe@ZnS/ZnS QD dispersion (optical density = 1.5 at 514 nm, corresponding to ∼25 mg/mL) in a mixed solvent of octane/hexane was deposited on the PEIE film and annealed at 150 °C for 5 min. Another PEIE EPL with a thickness of 7.5 nm (concentration of 0.5 wt% in 2-methoxyethanol) was then deposited by spin coating. Poly-TPD, 10 mg/mL in chlorobenzene, added as an HTL and MoOx, 10 mg/mL in acetonitrile, added as a hole injection layer (HIL) were each spin-coated at 3000 rpm for 30 s, followed by heating at 150 °C. All processes were performed in a glove box filled with nitrogen (H2O < 1 ppm). Finally, Al electrodes were deposited at a deposition rate of ∼6 Å/s by a thermal evaporation process under a high vacuum of ∼6 × 10⁻⁶ Torr.

2.3 Characterizations

UV-vis absorption spectra and PL emission spectra of the QD solution and films were obtained using a UV-visible spectrophotometer (JASCO, V630) and a Cary Eclipse fluorescence spectrometer (Agilent Technologies, G9803AA), respectively, at room temperature. The absolute PL QY of the QDs was measured using a quantum efficiency measurement system (Otsuka Electronics Co., Ltd., QD-2100). The surface roughness of the ZnO layer as a function of EBL PEIE thickness was measured using a high-resolution atomic force microscope (SII Nanotechnology, SPA-300HV). The images of the QDs and the cross-sectional images of the inverted QLEDs were obtained by high-resolution transmission electron microscopy (HR-TEM; JEOL, JEM-ARM 200F) at an accelerating voltage of 200 kV. The electrical and optical properties of the inverted QLEDs were determined using a spectroradiometer (Konica-Minolta, CS-2000) coupled with a voltage-current source unit (Tektronix, Keithley 2400). All the device evaluations were performed in air under ambient conditions and without encapsulation.

3. Results and discussion

First, with PEIE serving as the EPL, PL intensity was evaluated before and after rinsing the QD films with chlorobenzene at different thicknesses of the PEIE layer, as shown in Fig. 2(a). The PL intensity decreased until the PEIE EPL thickness was 5 nm, but was constant when the layer has a thickness of 7.5 nm, even after rinsing. This result indicates that the QD film is well preserved after rinsing with chlorobenzene. The thickness of the PEIE EPL needs to be minimized because of its insulating properties. Therefore, the optimal PEIE EPL thickness was determined to be 7.5 nm, based on the results obtained. The emission of QD films was monitored under 365 nm UV irradiation while dropping chlorobenzene onto the films to evaluate the damage of the EML upon HTL deposition, as shown in Fig. 2(b). A circular shape developed rapidly on the films without the PEIE EPL due to solvent evaporation from the surface, and a clear ring shape remained without many QDs left in the circle. In contrast, a less clear circle was observed on the films with 7.5-nm-thick PEIE EPL, indicating less damage by the chlorobenzene.

 

Fig. 2. (a) PL spectra of QD films depending on the PEIE thicknesses (without PEIE and with 5-nm- and 7.5-nm-thick layers) before and after chlorobenzene rinsing. (b) Photos after dropping chlorobenzene on QD films without and with a 7.5-nm-thick PEIE layer.

Download Full Size | PPT Slide | PDF

The integrity of the EML in the inverted QLED was further verified by cross-sectional TEM analysis, as shown in Fig. 3. A clear interface between the layers was observed with the 7.5-nm-thick PEIE EPL protected the EML even after HTL deposition, resulting in the formation of well-defined multilayers in the inverted QLED via the all-solution process. On the other hand, the QD EML without a PEIE EPL was considerably damaged and mixed with the adjacent poly-TPD. The QDs spread randomly and mixed with the MoOx HIL, leading to an uneven interface. A damaged bumpy surface is known to cause a non-uniform exciton recombination zone and leakage current in devices, leading to a decrease in device efficiency and stability [22–24]. The contact angle of the QD films with and without the PEIE EPL was measured, as shown in Figs. 3(c) and (d). The contact angle sharply decreased from 99.8° without the PEIE EPL to 63.5° with the 7.5-nm-thick PEIE EPL on the QD EML. Thus, we can conclude that the PEIE layer imparts hydrophilic properties on the QD EML and effectively protects the EML from HTL solvents.

 

Fig. 3. Cross-sectional TEM images of devices (a) without PEIE and (b) with a 7.5-nm-thick PEIE EPL. Contact angle measurement of the QD films (c) without and (d) with PEIE layer.

Download Full Size | PPT Slide | PDF

The PEIE EPL layer was applied to inverted QLEDs. The device performance was significantly improved by adding the 7.5-nm-thick PEIE EPL, as shown in Fig. 4. A maximum current efficiency (CE) of 31 cd A⁻¹ and EQE of 8.9% were obtained. These values indicate a six-fold increase in CE and EQE compared with those of devices without the PEIE layer. This significant improvement is attributed to the well-organized device structure discussed above. However, the overall excessive leakage current were observed in both devices, and the efficiency of the PEIE-EPL-based device was still relatively low, especially at low driving voltages.

 

Fig. 4. (a) Current density-voltage-luminance (J-V-L) characteristics. (b) CE and EQE as function of luminance of QLEDs with a 7.5-nm-thick PEIE EPL and without PEIE

Download Full Size | PPT Slide | PDF

A PEIE layer was then introduced as an EBL between the ZnO ETL and the QD EML to improve the charge balance in inverted QLEDs. It is known that excitons are typically quenched at the interface between the metal oxide and the QD EML due to the presence of surface defects of ZnO nanoparticles, resulting in a decrease in the PLQY of the QDs [10,12,30]. Therefore, the effects of the PEIE interlayer on exciton quenching were investigated by steady-state and time-resolved PL (TRPL) analysis of the QD films, as shown in Fig. 5. The PL intensity of the ZnO/QD film was significantly lower than that of the pristine QD film and the decrease in PL intensity was relatively less when a PEIE EBL was added on the ZnO film. This result indicates that some of the excitons generated from the QDs were extinguished by ZnO film and the PEIE EBL effectively suppresses exciton quenching. The PL decay time was measured and fitted to a biexponential decay model. The average PL decay time of the ZnO/QD film increased from 5.6 ns to 7.2 ns with a 5-nm-thick PEIE EBL on the ZnO film, which is in good agreement with the above results. PEIE is believed to reduce non-radiative exciton quenching by preventing the direct contact of the QD EML with the surface defects of the ZnO film.

 

Fig. 5. (a) Steady-state PL spectra of glass/QD, glass/ZnO/QD, and glass/ZnO/PEIE/QD films. (b) Time-resolved PL of glass/QD, glass/ZnO/QD, glass/ZnO/PEIE/QD films, and QDs in solution

Download Full Size | PPT Slide | PDF

Next, the surface roughness of the ZnO film was investigated with different PEIE EPL thicknesses, as shown in Fig. 6. Rough surfaces are known to cause high contact resistance and leakage current, leading to low efficiency and short lifetime of QLEDs [22,38]. The roughness of the ZnO films was gradually reduced from 1.37 nm (pristine ZnO film) to 0.52 nm with the increase in PEIE EPL thickness. The thin PEIE EPL reduced the RMS roughness of the ZnO surface by filling the deep valleys of the ZnO film. This flat film morphology enables uniform QD EML deposition and reduces leakage current [38].

 

Fig. 6. AFM images of ZnO film with different PEIE layer thicknesses; (a) without PEIE and (b) PEIE layers with thicknesses of 2 nm, (c) 5 nm, (d) 7.5 nm. (An area of 5 × 5 μm2 was measured for AFM)

Download Full Size | PPT Slide | PDF

The inverted QLEDs were fabricated with the PEIE EBL, and their electroluminescence (EL) properties and performance were characterized, as shown in Fig. 7 and Table 1. The layer structure and a flat-band energy diagram of an inverted QLED with dual PEIE interfacial layer are illustrated in Figs. 7(a) and (b). The EL emission spectrum exhibits a peak at 524 nm with a FWHM of 21 nm, which is considered a highly pure green emission (as expressed in the Commission Internationale de l'Eclairage (CIE) color coordination of (0.14, 0.79)). No change in the EL emission spectra was observed with different PEIE EBL thicknesses. As shown in current density-voltage-luminance (J-V-L) characteristics, the current density gradually decreased, especially below turn-on voltage, with the increase in PEIE EBL thickness. This is attributed to the smooth surface and reduced leakage current. Although PEIE is an insulating polymer, the current density remained similar above the turn-on voltage, indicating that a PEIE layer, with up to 5 nm thickness, does not affect electron injection. The maximum luminance was enhanced by a factor greater than 2 when a 5-nm-thick PEIE EBL was added. The device efficiency further improved with the addition of the PEIE EBL and gradually increased until the PEIE EBL thickness was 5 nm. In particular, the efficiency significantly improved at low current densities due to the reduced leakage current, as shown in Fig. 7(f). A maximum CE of 54.1 cd A⁻¹ and EQE of 12.4% were obtained for the device with the 5-nm-thick PEIE EBL. These values are approximately 1.6 times higher than those for devices without a PEIE EBL. On the other hand, the turn-on voltage slightly increased and the efficiency decreased for the device with the 7.5-nm-thick PEIE EBL, indicating that a PEIE layer that is too thick limits electron injection.

 

Fig. 7. (a) Schematic structure and (b) flat-band energy level diagram of the all-solution-processed inverted QLED. (c) CIE coordinates (0.14, 0.79) marked with a star. (d) Normalized EL spectra of the QLEDs with different PEIE EBL thicknesses. Device characteristics of inverted QLEDs with different PEIE EBL thicknesses: (e) Current density-voltage-luminance (J-V-L) characteristics. (f) CE and EQE as a function of luminance of QLEDs.

Download Full Size | PPT Slide | PDF

Table 1. Device performance of QLEDs with different PEIE layer thicknesses.

View Table

Single carrier devices were fabricated to understand charge transport through the PEIE EBL, as shown in Fig. 8. The current density of the electron-only device (EOD) without PEIE is more than three orders of magnitude higher than that of the hole-only device (HOD). However, the current density of the EOD decreases gradually as the PEIE EBL thickness increases, which indicates that insulating property of PEIE EBL mitigate excessive electron injection to the QD EML, and the leakage current of the device is mainly due to electron leakage. We believe that the hole and electron carrier densities are relatively balanced in a 5 nm-thick PEIE EBL, which results in the most improved device efficiency.

 

Fig. 8. Current densities of hole-only device (HOD) and electron-only device (EOD) with different PEIE EBL thicknesses. (HOD: ITO/poly-TPD/QD/PEIE/poly-PTD/MoOx/Al, EOD: ITO/ZnO/PEIE/QD/ZnO/Al)

Download Full Size | PPT Slide | PDF

We fabricated over 20 devices to confirm the reproducibility of the resulting QLEDs. The maximum CE, EQE, and the maximum luminance of the device with the 5-nm-thick PEIE EBL reached 70.0 cd A⁻¹, 17.3%, and 380 000 cd m⁻², respectively. The luminance obtained in this study is the highest value reported for all-solution-processed inverted QLEDs investigated so far. Thus, we can conclude that the PEIE EBL not only reduces exciton quenching caused by ZnO but also suppresses excessive electron injection to the QD EML, leading to the enhancement of device performance.

4. Conclusion

Quantum dot (QD) emitting layers (EML) were sandwiched between polyethylenimine-ethoxylated (PEIE) layers for all-solution-processed inverted quantum dot light-emitting diodes (QLEDs). First, PEIE was coated onto the QD EML as an EML protecting layer (EPL), and the PEIE thickness was optimized by investigating solvent resistance and device integrity. The hydrophilic property of PEIE effectively protects the QD EML from the hydrophobic HTL solvent damage. The optimal PEIE EPL thickness was determined to be 7.5 nm, which is close to the minimum thickness required to protect the QD EML. A PEIE layer was then introduced as an electron-blocking layer (EBL) between the ZnO electron ETL and QD EML to improve the charge balance in the inverted QLEDs. The insulating property of PEIE effectively suppressed the electron injection to the QD EML. The inverted QLEDs were thus fabricated to examine the effect of the PEIE EPL, and their performances were compared to those without PEIE layers. The maximum current efficiency increased from 31.1 cd A⁻¹ in a device without the PEIE EBL to 54.1 cd A⁻¹ in a device with a 5-nm-thick PEIE EBL. The maximum external quantum efficiency (EQE) was also enhanced from 8.9% to 12.4%. In addition, the maximum luminance was drastically improved by a factor of 2.3 with the introduction of the PEIE EBL. The 5-nm-thick PEIE EBL not only reduced exciton quenching caused by ZnO but also suppressed the excessive electron injection to the QD EML. This study demonstrated that the QD EML structure sandwiched with PEIE layers is an effective approach for fabricating all-solution-processed inverted QLEDs.