Enhancing the Performance of Quantum Dot Light-Emitting Diodes Using Solution-Processable Highly Conductive Spinel Structure CuCo2O4 Hole Injection Layer

Charge imbalance in quantum-dot light-emitting diodes (QLEDs) causes emission degradation. Therefore, many studies focused on improving hole injection into the QLEDs-emitting layer owing to lower hole conductivity compared to electron conductivity. Herein, CuCo2O4 has a relatively higher hole conductivity than other binary oxides and can induce an improved charge balance. As the annealing temperature decreases, the valence band maximum (VBM) of CuCo2O4 shifts away from the Fermi energy level (EF), resulting in an enhanced hole injection through better energy level alignment with hole transport layer. The maximum luminance and current efficiency of the CuCo2O4 hole injection layer (HIL) of the QLED were measured as 93,607 cd/m2 and 11.14 cd/A, respectively, resulting in a 656% improvement in luminous performance of QLEDs compared to conventional metal oxide HIL-based QLEDs. These results demonstrate that the electrical properties of CuCo2O4 can be improved by adjusting the annealing temperature, suggesting that solution-processed spinel can be applied in various optoelectronic devices.


Introduction
Quantum dots (QDs) have a size-dependent bandgap because of the quantum confinement effect; thus, the emission wavelength of QDs can be easily controlled using this property [1]. Inorganic QDs have narrower electroluminescence (EL) spectra and higher quantum yield (QY) than conventional organic light-emitting materials [2,3]. In addition, QDs have the advantage of low-cost mass production of quantum-dot light-emitting diodes (QLEDs) through solution-based manufacturing technology [4]. Owing to these advantages, QDs are in the spotlight as light-emitting layers for next-generation displays.
To increase the emission performance of QLED, hole and electron injection/transport layers are inserted between the electrodes and the QD emitting layer (EML) [5]. Various studies are being conducted to employ transition metal oxides (TMOs) as the charge transport layer of QLEDs because TMOs have strong moisture and heat resistance. In particular, these studies applied p-type TMOs, such as CuO and NiO, as hole injection layers (HIL) or hole transport layers (HTL) [6][7][8], and n-type TMOs, such as ZnO and ZnMgO, as electron transport layers (ETL) [9,10] in optoelectronic devices.

CuCo 2 O 4 Solution Synthesis
First, 0.48 mmol of copper acetate tetrahydrate (Cu(CH 3 CO 2 ) 2 ·H 2 O, Sigma Aldrich) and 0.96 mmol of cobalt acetate tetrahydrate (Co(CH 3 CO 2 ) 2 ·4H 2 O, DAE JUNG) were dissolved in 12 mL ethylene glycol monomethyl ether (DAE JUNG). The solution was then sonicated for 10 min at room temperature to completely dissolve the copper and cobalt precursors. Subsequently, 60 µL of ethanolamine (Sigma Aldrich, St. Louis, MI, USA) was added and stabilized by hydrogen bonding using a stabilizer and stirred for 24 h at room temperature. After stirring for 24 h, the 0.04 M CuCo 2 O 4 solution changed from blue to dark gray (or dark navy).

Device Fabrication
Patterned indium tin oxide (ITO) glass substrates were cleaned via ultrasonication in DI water, acetone, and isopropyl alcohol for 15 min in sequence. The cleaned ITO substrates were treated with ultraviolet-ozone for 20 min to improve the surface hydrophilicity and increase the ITO work function through carbon contamination removal. The CuCo 2 O 4 HIL solution was spin-coated at 3000 rpm for 60 s. Subsequently, the solvent was dried on a hot plate at 150 • C for 10 min and then annealed for 60 min (200 • C/300 • C) in an ambient atmosphere. As shown in Equations (1) and (2), acetate residues are removed in the pre-annealing and CuCo 2 O 4 is formed through post-annealing. Cu(CH 3 CO 2 ) 2 + 2Co(CH 3 CO 2 ) 2 + 6H 2 O → Cu(OH) 2 + 2Co(OH) 2 + 6 Acetic acid (1) Cu(OH) 2 + 2Co(OH) 2 A poly[(9,9-dioctylfluorenyl-2,7-diyl)-co-(4,4 -(N-(4-sec-butylphenyl)diphenylamine)] (TFB) dissolved in p-xylene at 1 wt% was spin-coated at 3000 rpm for 30 s onto the CuCo 2 O 4 HIL, followed by annealing at 180 • C for 40 min. Then, CdSe/ZnS Green QDs (UNIAM, 20 mg/mL) dispersed in toluene were spin-coated at 2000 rpm for 30 s and then annealed at 90 • C for 10 min. The optical bandgap and photoluminescence spectra of the green QDs used to fabricate the QLEDs are shown in Figure S2. ZnO (Avantama, N-10) dispersed in isopropyl alcohol was spin-coated at 2000 rpm for 60 s and then annealed at 90 • C for 10 min. Finally, a 130 nm thick aluminum was deposited by a thermal evaporation at a deposition rate of 3 Å/s under high vacuum through a shadow mask.
A HOD and an electron-only device (EOD) to evaluate the charge injection ability were produced through spin coating-based solution process identical to the QLED manufacturing process. First, pre-cleaned glass/ITO substrate were treated with ultraviolet-ozone for 20 min. The CuCo 2 O 4 HIL solution was spin coated at 3000 rpm for 60 s. Subsequently, pre-annealing is performed on the hot plate at 150 • C for 10 min and post-annealing is performed annealed for 60 min at 200 • C (or 300 • C) in an ambient atmosphere. Next, TFB HTL solution was spin-coated at 3000 rpm for 30 s onto the CuCo 2 O 4 HIL, followed by annealing at 180 • C for 40 min. Finally, a 130 nm thick aluminum was deposited by a thermal evaporation at a deposition rate of 3 Å/s under high vacuum through a shadow mask. First, pre-cleaned glass/ITO substrate were treated with ultraviolet-ozone for 20 min. ZnO ETL was spin-coated at 2000 rpm for 60 s and then annealed at 90 • C for 10 min. Then, a 130 nm thick aluminum was deposited by a thermal evaporation at a deposition rate of 3 Å/s under high vacuum through a shadow mask.

Characterization
The electrical properties such as resistivity, carrier mobility, and carrier concentration of the oxide thin film were measured using Hall measurements (HL 5500PC). XPS and UPS spectra were measured using a surface analysis system (Thermo fisher, NEXSA, Waltham, MA, USA) with an Al Kα (1486.6 eV) source for XPS and a He I (21.22 eV) source for UPS. The energy references of the XPS and UPS spectra were calibrated with respect to the E F of clean Au sample. The morphology of the films was characterized using atomic force microscopy (AFM) (S.I.S-GmbH, Berlin, Germany) in the non-contact mode. The transmittance of the thin films was measured using a UV-visible spectrometer (Cary 100, Agilent). The electroluminescence properties of the QLEDs were measured using an I-V-L system (M-6100, McScience) installed with a source meter (Keithley 2400) and spectroradiometer (CS-2000, Konica Minolta). QLED cross-section images were obtained using HR-TEM, and the elemental composition of the cross-sections was analyzed using energy-dispersive spectroscopy (EDS) data using a field emission electron microscope (JEM-2100F, JEOL).    Figure 1 displays the O 1s, Cu 2p 3/2 and Co 2p 3/2 core-level XPS spectra of ITO/CuCo 2 O 4 at different annealing temperatures of 200 • C and 300 • C. The binding energies of the XPS spectra were calibrated using the C 1s peak of adventitious carbon (284.6 eV as a reference). The O 1s core level was separated into oxide peaks, oxygen vacancies, surface hydroxyl oxygen, and carbonyl oxygen. As seen in Figure 1a,b, more oxygen vacancies and hydroxyl oxygen peaks were observed in the CuCo 2 O 4 -200 specimen than in the CuCo 2 O 4 -300 specimen, which means that more oxygen defects occurred. The tendency of oxygen defects in the CuCo 2 O 4 to increase at the lower annealing temperature is similar to reported solution-processed metal oxides [29,30]. Cu + ions do not cause satellite peaks because the 3d orbitals are all filled with electrons, whereas Cu 2+ ions have [Ar] 3d 9 electron structures, which cause 2p→3d transitions with the result that satellite peaks are observed in XPS measurement [31][32][33][34]. As seen in Figure 1c,d, two satellite peaks were identified at 940 eV and 943 eV, except for the main peak at 933 eV in the Cu 2p 3/2 core level. Therefore, the Cu 2p 3/2 XPS spectra can be separated into Cu + , Cu 2+ , Cu(OH) 2 , and two satellite peaks (Sat1 and Sat2), and the ratio of Cu + to Cu 2+ can be expressed as Equations (4) and (5), respectively [19,35,36].
In the case of the CuCo 2 O 4 -200 and CuCo 2 O 4 -300, the Co 3+ ion ratios were 70.9% and 72.3%, respectively. In p-type metal oxides, oxygen vacancies reduce the number of hole carriers, which has a significant influence on the electrical properties of the oxides [12,38]. The XPS measurement indicates that the solution-processed CuCo 2 O 4 has no significant change in Cu and Co metal ion composition according to the annealing temperature change; nevertheless, there is a significant change in the oxygen defect. The O 1s peak measurements show that more oxygen defect sites were generated by low-temperature annealing in CuCo 2 O 4 -200. Therefore, we can expect that the hole concentration in CuCo 2 O 4 -200 is less than that in CuCo 2 O 4 -300. These results show the same tendency as the Hall measurement results, in which CuCo 2 O 4 -200 exhibited a lower hole concentration. Figure 2 shows an AFM topography image of the CuCo 2 O 4 thin films with different annealing temperatures measured in the non-contact mode. As shown in Figure 2a, the CuCo 2 O 4 -200 thin film has low crystallinity, whereas the CuCo 2 O 4 -300 thin film in Figure 2b shows that both the size and number of particles increased. Root mean square (RMS) value of CuCo 2 O 4 -200 and -300 thin films were measured to be 0.17 nm and 0.46 nm, respectively. The uniformity of the film formation seems to be the reason why the CuCo 2 O 4 -200 thin film had a higher hole mobility than CuCo 2 O 4 -300 (Table 1). As the CuCo 2 O 4 annealing temperature decreased, crystallinity also decreased. Therefore, the low crystallinity of CuCo 2 O 4 -200 is a major factor in increasing the number of oxygen defect sites. Co 2+ and two satellite peaks (Sat 1, Sat 2); the ratio for each ion can be expressed by Equations (6) and (7), respectively [37].
Co (%) = Co + Sat1 Co + Co + Sat1 + Sat2 * 100 (6) Co (%) = Co + Sat2 Co + Co + Sat1 + Sat2 * 100 (7) In the case of the CuCo2O4-200 and CuCo2O4-300, the Co 3+ ion ratios were 70.9% and 72.3%, respectively. In p-type metal oxides, oxygen vacancies reduce the number of hole carriers, which has a significant influence on the electrical properties of the oxides [12,38]. The XPS measurement indicates that the solution-processed CuCo2O4 has no significant change in Cu and Co metal ion composition according to the annealing temperature change; nevertheless, there is a significant change in the oxygen defect. The O 1s peak measurements show that more oxygen defect sites were generated by low-temperature annealing in CuCo2O4-200. Therefore, we can expect that the hole concentration in CuCo2O4-200 is less than that in CuCo2O4-300. These results show the same tendency as the Hall measurement results, in which CuCo2O4-200 exhibited a lower hole concentration. Figure 2 shows an AFM topography image of the CuCo2O4 thin films with different annealing temperatures measured in the non-contact mode. As shown in Figure 2a, the CuCo2O4-200 thin film has low crystallinity, whereas the CuCo2O4-300 thin film in Figure  2b shows that both the size and number of particles increased. Root mean square (RMS) value of CuCo2O4-200 and -300 thin films were measured to be 0.17 nm and 0.46 nm, respectively. The uniformity of the film formation seems to be the reason why the CuCo2O4-200 thin film had a higher hole mobility than CuCo2O4-300 ( Table 1). As the CuCo2O4 annealing temperature decreased, crystallinity also decreased. Therefore, the low crystallinity of CuCo2O4-200 is a major factor in increasing the number of oxygen defect sites.  The optical characteristics of the CuCo 2 O 4 -200 and -300 films were measured by UV-visible spectroscopy. As shown in Figure 3a, the transmittance (%) of the CuCo 2 O 4 films decreased as the annealing temperature was increased. Especially, the transmittance of CuCo 2 O 4 -200 and -300 samples was measured to be 91.1% and 88.4% respectively, in the 530 nm region, which is the QLED emission wavelength. The Tauc plots for the CuCo 2 O 4 thin films annealed at different temperatures are shown in Figure 3b. The spinel oxides include divalent metal ions (M 2+ ) and trivalent metal ions (M 3+ ), and, thus, conduction bands formed by the 3d orbitals of M 2+ and 3d orbitals of M 3+ exist, respectively. Because both O 2p→M 2+ 3d and O 2p→M 3+ 3d transitions are generated, spinel oxides have two bandgap energies [39][40][41][42]. E g1 and E g2 in Figure 3b are the optical bandgaps generated by the O 2p→M 2+ 3d and O 2p→M 3+ transitions, respectively. The optical bandgap of CuCo 2 O 4 -200 with spinel structure was measured at 1.36 eV and 2.15 eV, and the bandgap of CuCo 2 O 4 -300 was measured at 1.32 eV and 2.13 eV. According to prior studies on spinel cobalt oxide, the crystallinity increases as the annealing temperature increases, whereas the bandgap and transmittance will decrease. The optical characterization of the CuCo 2 O 4 covered in this study showed the same tendency as previous studies on spinel oxides [43]. Through UV-visible spectroscopy measurement results, owing to the low crystallinity of CuCo 2 O 4 -200, it was confirmed that the CuCo 2 O 4 -200 transmittance was higher than that of CuCo 2 O 4 -300 and, consequently, more suitable for QLED. The optical characteristics of the CuCo2O4-200 and -300 films were measured by UVvisible spectroscopy. As shown in Figure 3a, the transmittance (%) of the CuCo2O4 films decreased as the annealing temperature was increased. Especially, the transmittance of CuCo2O4-200 and -300 samples was measured to be 91.1% and 88.4% respectively, in the 530 nm region, which is the QLED emission wavelength. The Tauc plots for the CuCo2O4 thin films annealed at different temperatures are shown in Figure 3b. The spinel oxides include divalent metal ions (M 2+ ) and trivalent metal ions (M 3+ ), and, thus, conduction bands formed by the 3d orbitals of M 2+ and 3d orbitals of M 3+ exist, respectively. Because both O 2p→M 2+ 3d and O 2p→M 3+ 3d transitions are generated, spinel oxides have two bandgap energies [39][40][41][42]. Eg1 and Eg2 in Figure 3b are the optical bandgaps generated by the O 2p→M 2+ 3d and O 2p→M 3+ transitions, respectively. The optical bandgap of CuCo2O4-200 with spinel structure was measured at 1.36 eV and 2.15 eV, and the bandgap of CuCo2O4-300 was measured at 1.32 eV and 2.13 eV. According to prior studies on spinel cobalt oxide, the crystallinity increases as the annealing temperature increases, whereas the bandgap and transmittance will decrease. The optical characterization of the CuCo2O4 covered in this study showed the same tendency as previous studies on spinel oxides [43]. Through UV-visible spectroscopy measurement results, owing to the low crystallinity of CuCo2O4-200, it was confirmed that the CuCo2O4-200 transmittance was higher than that of CuCo2O4-300 and, consequently, more suitable for QLED.    Figure 4a, the graph on the left shows the work function measurement based on the kinetic energy of electrons in the secondary-electron cutoff region. In Figure 4a, the graph on the right shows the energy difference between the Fermi energy level (E F ) and the VBM or highest occupied molecular orbital ( Figure S3). The work function difference and HOMO level difference between the TFB thin film formed on CuCo 2 O 4 -200 and formed on CuCo 2 O 4 -300 were 0.01 eV and 0.02 eV, respectively, and these were not significant. However, in the CuCo 2 O 4 HILs, an energy level difference appeared when the annealing temperature changed, and the VBM in particular changed significantly. The CuCo 2 O 4 -300 HIL has few oxygen defect sites owing to its high crystallinity, which increases the hole concentration and correspondingly reduces the gap between the VBM and E F levels. Contrastingly, the CuCo2O4-200 HIL has many oxygen defect sites, which decreases the hole concentration and correspondingly increases the gap between the VBM and E F levels. As shown in Figure 4b, the difference between the CuCo 2 O 4 -300 VBM and TFB HOMO levels was 0.7 eV, whereas the difference between CuCo 2 O 4 -200 and TFB HOMO levels was 0.49 eV, which is relatively small. The UPS measurement results suggested that the CuCo 2 O 4 -200/TFB structure was more suitable for the device in terms of hole injection.
To confirm the degree of hole injection in actual devices according to the CuCo 2 O 4 HILs annealing temperature, the J-V characteristics were measured by manufacturing an HOD of the ITO/CuCo 2 O 4 /TFB/Al structure, as displayed in Figure 5a. An EOD with an ITO/ZnO/Al structure was also fabricated to compare the injection and transport properties of the electrons and hole carriers. Figure 5b shows that more current flows in the EOD than in the HOD, which can be expected based on the characteristics of holes and electrons. Remarkably, the current density of the HODs using the ternary metal oxide CuCo 2 O 4 HIL was nearly 100 times higher than that of the binary metal oxide CuO. In addition, as expected from the UPS measurement results, it can be confirmed that hole injection increases in the CuCo 2 O 4 -200 HOD, which exhibits suitable energy alignment with the TFB. Therefore, we expect that the CuCo 2 O 4 HIL can improve the performance of QLEDs owing to its high hole injection ability, and in particular, that CuCo 2 O 4 -200 HIL QLEDs will have the highest performance. Figure 6a shows the structure of the QLED that was fabricated using the solution process. First, the CuCo 2 O 4 , TFB, QDs, and ZnO layers were prepared by spin-coating on a patterned glass/ITO substrate. Then, the 130 nm Al cathode was thermally evaporated using a metal shadow mask. Figure 6b shows a cross-sectional HR-TEM image of the QLED fabricated using the solution process. The HR-TEM image shows that CuCo 2 O 4 /TFB/QDs/ZnO thin films were formed with uniform thickness. The thicknesses of the CuCo 2 O 4 -200, TFB, QD, and ZnO layers in the TEM images were 8, 13, 40, and 35 nm, respectively. EDS line scanning (Figure 6c) displayed the overall elemental distribution, confirming that each solution-processed thin film was well separated. The EDS mapping images of C, O, S, In, Co, Cu, and Zn are shown in Figure S4. In the EDS data, copper (Cu) and cobalt (Co) peaks are co-located, indicating that Cu and Co do not separate and form a monolayer. In particular, the atomic ratio of the CuCo 2 O 4 layer analyzed by EDS was Cu:Co:O = 0.9:2:4.1 ( Figure S5). This atomic ratio provides evidence of the formation of a stable spinel (AB 2 O 4 ) structure. To confirm the degree of hole injection in actual devices according to the HILs annealing temperature, the J-V characteristics were measured by manufact HOD of the ITO/CuCo2O4/TFB/Al structure, as displayed in Figure 5a. An EOD ITO/ZnO/Al structure was also fabricated to compare the injection and transpor of QLEDs owing to its high hole injection ability, and in particular, that CuCo2O QLEDs will have the highest performance.  Figure 6a shows the structure of the QLED that was fabricated using th process. First, the CuCo2O4, TFB, QDs, and ZnO layers were prepared by spina patterned glass/ITO substrate. Then, the 130 nm Al cathode was thermally e using a metal shadow mask. Figure 6b shows a cross-sectional HR-TEM im QLED fabricated using the solution process. The HR-TEM image s CuCo2O4/TFB/QDs/ZnO thin films were formed with uniform thickness. The t of the CuCo2O4-200, TFB, QD, and ZnO layers in the TEM images were 8, 13, nm, respectively. EDS line scanning (Figure 6c) displayed the overall element tion, confirming that each solution-processed thin film was well separated. The ping images of C, O, S, In, Co, Cu, and Zn are shown in Figure S4. In the EDS d (Cu) and cobalt (Co) peaks are co-located, indicating that Cu and Co do not se form a monolayer. In particular, the atomic ratio of the CuCo2O4 layer analyz   Table 2. As shown in Figure 7, the current density at low voltage of the CuCo 2 O 4 HIL QLEDs was higher than that of the CuO HIL QLED. However, the current density of the CuO HIL QLED rapidly increased after turning on the device, whereas the CuCo 2 O 4 -based device showed a gradual increase rate of the current density due to a decrease in the leakage current of the device. Therefore, the current efficiency was higher for QLEDs with CuCo 2 O 4 HIL than CuO HIL and a narrow FWHM could be realized, allowing higher color purity of the device. This is due to the enhanced charge balance of the device with better hole injection. This study demonstrates that efficient hole injection can be achieved using a solution-processed spinel CuCo 2 O 4 HIL in QLEDs.  Figure 7 shows the light-emitting performance of the CuCo2O4 and CuO-HIL-based QLEDs. Figure 7a shows the current density-voltage-luminance (J-V-L) characteristics of QLEDs, and all CuCo2O4 HIL-based QLEDs show a higher luminance than the CuO HILbased QLED. The performance improvement of the QLEDs is owing to the excellent hole injection characteristics of CuCo2O4 and shows the same tendency as the J-V characteris- with CuCo2O4 HIL than CuO HIL and a narrow FWHM could be re color purity of the device. This is due to the enhanced charge bal better hole injection. This study demonstrates that efficient hole in using a solution-processed spinel CuCo2O4 HIL in QLEDs.

Conclusions
In this study, we fabricated CuCo 2 O 4 HIL-based QLEDs to increase hole injection and improve luminous efficiency. The electrical properties of CuCo 2 O 4 comprising Cu 2+ and Co 3+ cations were measured using Hall measurements, and it was confirmed that they have higher conductivity than common p-type binary oxides. In particular, the difference in the hole injection ability of CuCo 2 O 4 according to the annealing temperature was identified using XPS, UPS, and the fabricated HOD. It was confirmed that the VBM of CuCo 2 O 4 shifted away from the Fermi energy level (E F ) as the annealing temperature decreased, resulting in enhanced hole injection through better energy-level alignment. The CuCo 2 O 4 HODs had a significantly higher current density than the CuO HOD, and in particular, the hole injection ability of the CuCo 2 O 4 -200 HOD was superior to that of the CuCo 2 O 4 -300 HOD, as expected from the perspective of energy alignment. Maximum luminance and maximum current efficiency of the CuCo 2 O 4 -200 HIL QLED were 93,607 cd/m 2 and 11.14 cd/A, respectively, resulting in a 656% improvement in luminous performance of QLEDs compared to the CuO HIL QLED owing to increased hole injection. This study demonstrates that efficient hole injection can be achieved using a solution-processed spinel CuCo 2 O 4 HIL in QLEDs.

Supplementary Materials:
The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/ma16030972/s1, Figure S1. Crystallographic structure of AB2O4 spinel. Figure S2. (a) The optical band gap of CdSe/ZnS green QDs. (b) The Photoluminescence spectra of CdSe/ZnS green QDs. Figure S3. The optical band gap of TFB was estimated using the Tauc plot. Figure S4. EDS mapping image of various elements (C, O, S, In, Co, Cu, Zn and Se) in the CuCo2O4 HIL based-QLED. Figure S5. EDS line scan data of atomic percent. Figure S6. Normalized EL spectra of QLED at the maximum luminance.  Data Availability Statement: Data will be made available from the corresponding authors on reasonable request.

Conflicts of Interest:
The authors declare no conflict of interest.