1. Introduction

Colloidal quantum dots (QDs) have attracted much attention in both research and industry fields due to their unique properties, such as high photoluminescence (PL) quantum efficiency, tunable and saturated emission color, low-cost solution processability and potential applications in optoelectronic devices [1–6]. The applications of QDs as emissive layer excited by light in phosphors converted light emitting diodes (LEDs) or driven by electrical current in sandwiched electroluminescence (EL) devices have been well demonstrated. Especially, the light emitting diodes based on QDs as the emission layer (QLEDs) are considered as the alternative coordinates or complement technology for the next generation lighting and display technologies [7–10]. In 1994, the first QLED was reported by A.P. Alivisatos [11]. Since then, much progress has been achieved by optimizing both the QD and device structures. For the material aspect, the QDs with thick shell were successfully synthesized and used as the emission layer, which is a rather important progress and prerequisite to construct efficient QLEDs [12–16]. On the other hand, the hybrid QLEDs with ZnO as the electron injection layer (EIL) show very outstanding performance, including high luminance of over 200000 cd/m² and external quantum efficiency of 20.5% [17,18]. It is widely accepted that the introduction of ZnO EIL is considered as a milestone for the QLEDs. To date, the performance of solution processed Cd based QLEDs at the lab scale are comparable to state-of-the-art organic light emitting diodes fabricated by vacuum deposition. Moreover, compared to organic light emitting diodes, QLEDs have many prominent advantages, such as low-cost solution process ability and wider color gamut. However, the industrial technology maturity of QLEDs still lags behind that of organic light emitting diodes and it is desired to further decrease the production cost of QLEDs to meet the request of commercialization. Reducing the cost of raw materials used in QLEDs is one of a feasible and effective way.

The common used ZnO EIL is obtained by spin-coating a ZnO nanoparticle (n-ZnO) ethanol solution and the n-ZnO is synthesized by a method as reported by other groups [19,20] As we know, the solution synthesis processes are high-cost and time-consuming, which increases the manufacturing costs of QLED inevitably. Hence, a low-cost and timesaving method is desired for the EIL in QLEDs. We know that the ZnO film can also be obtained by a sol-gel ZnO precursor solution, which is often employed in solar cells [21–23]. The sol-gel ZnO (s-ZnO) precursor solution is obtained by mixing the zinc acetate dehydrate and ethanolamine in ethanol solvent, which is easily scaled up for industrial production.

In this study, inverted devices structure consisting of ITO/s-ZnO/QDs/ 4,4'-Bis(carbazol-9-yl)biphenyl (CBP)/MoO₃/Al was built. In addition, the device with n-ZnO as EIL was also fabricated as a reference. The results demonstrate that the s-ZnO is competent for QLEDs instead of common used n-ZnO. The performance of s-ZnO-containing device is comparable to that of the device based on n-ZnO EIL. It is worthy to note that the s-ZnO film is obtained by a low-cost sol-gel ZnO solution, which is superior to that of n-ZnO achieved through solution synthesis and high-speed centrifugation processes.

2. Experimental details

The synthesis methods for QDs and n-ZnO were depicted in the literature [19]. The concentration of n-ZnO is 30 mg/ml. For s-ZnO, the ZnO precursor solution was prepared by dissolving 0.5 M zinc acetate and 0.5 M monoethanolamine in 2-methoxyethanol with vigorous stirring for 12 h for the hydrolysis reaction at ambient conditions. The sol–gel-derived ZnO layer was prepared by spin-coating the ZnO precursor solution onto an UV–ozone treated ITO substrate at a rate of 4000 rpm for 40 s, and then thermally annealing at 150 °C for 5 min in ambient air, during which the precursor was converted to a dense ZnO film by hydrolysis. The devices with a standard structure of ITO/EIL (45 nm)/QDs (30 nm) /CBP (55 nm)/MoO₃ (7 nm)/Al (150 nm) were fabricated with solution processes for the EIL and QD layer. The CBP, MoO₃ and Al layers were deposited by thermal evaporation technology. The n-ZnO films were deposited onto the UV-ozone treated ITO substrates at 2000 rpm for 60 seconds and then annealed at 100°C for 30 min. The QDs were dispersed in toluene (5 mg/ml) and spin-coated onto the two kinds of ZnO layers at 2500 rpm and then annealed at 80 °C for 30 min. Finally, these substrates were transferred into a chamber to finish the fabrication of CBP, MoO₃ and aluminum (Al) layers at a pressure of 4.5 × 10⁻⁴ Pa. The layer thicknesses were 60, 8, and 100 nm for CBP, MoO₃, and Al, respectively.

The EL spectra of the devices were measured by a PR655 spectrometer, and the current density (J)–voltage (V)–luminance (L) characteristics were recorded simultaneously by combining the spectrometer with Keithley model 2400 programmable voltage–current source. Absorption spectra of the samples were recorded by an UV/vis/near-IR spectrophotometer (Perkin Elmer Lambda 750) at room temperature. The steady state photoluminescence (PL) spectra were performed with Fluorescence spectrometer (Horiba Jobin Yvon Flworomax-4). The surface morphology of the ITO with NP-ZnO and Sol-Gel-ZnO were characterized by Atomic force microscope (AFM) (Veeco MultiMode V) and scanning electron microscope (SEM) (FEI, Quanta 200FEG). Elemental composition and the electronic structures of the film were determined by ultraviolet photoelectron spectroscopies (UPS) using a Kratos AXIS Ultra-DLD ultrahigh vacuum (UHV) surface analysis system with a monochromatic aluminum Ka source (1486.6 eV) and UPS analysis were carried out with an unfiltered HeI (21.2 eV) gas discharge lamp and a hemispherical analyzer. All spectra were measured at room temperature.

3. Results and discussion

Figure 1 shows the SEM images of n-ZnO, s-ZnO layers, as well as the QDs films on both types of ZnO layers. The samples are consisted of glass/ITO/n-ZnO, glass/ITO/n-ZnO/QDs, glass/ITO/s-ZnO, and glass/ITO/s-ZnO/QDs. We can see that a smooth surface can be achieved for both the s-ZnO and n-ZnO seen from Fig. 1(a) and 1(c). Figure 1(b) and 1(d) show the SEM images of QD on these two types of ZnO layers. It can be seen that both of the QD layers exhibit a close packed structure with only a few cracks. These results demonstrate that the s-ZnO is as good as the common used n-ZnO for inverted QLEDs from the morphology property side.

 

Fig. 1 SEM images for (a) n-ZnO, (b) n-ZnO /QDs, (c) s-ZnO, and (d) s-ZnO /QDs.

Download Full Size | PDF

To further investigate the quality of each layer, we carried out the AFM measurements for the samples used in the SEM measurement. As shown in Fig. 2(a) and 2(c), the s-ZnO film shows smoother surface than that of n-ZnO layer on the glass/ITO substrates, and the corresponding root-mean-square (RMS) values are 3.92 and 2.25 nm for n-ZnO and s-ZnO, respectively. As we know, a smooth surface of QD layer plays a rather important role in determining the device performance, especially the electrical properties. In other words, the smoother surfaces can prevent electrical shorts or formation of preferred current channels through QLEDs and benefit to device performance. Figure 2(b) and 2(d) show the AFM image of QD layers on the two ZnO films. The RMS values are 1.97 and 1.50 nm for QD layers on n-ZnO and s-ZnO, respectively. These results can further demonstrate that high quality and smooth surface of QD film can be achieved on s-ZnO EIL layer, which comparable to that of QD film on n- ZnO EIL layer.

 

Fig. 2 AFM images for (a) n-ZnO, (b) n-ZnO /QDs, (c) s-ZnO, and (d) s-ZnO /QDs.

Download Full Size | PDF

To access the availability of s-ZnO as the EIL in inverted QLEDs, we fabricated two QLEDs with s-ZnO and n-ZnO as the EILs. In detailed, the QD emissive layer is sandwiched between the n-ZnO EIL and CBP hole transport layer (HTL) for the control device. And the s-ZnO instead of n-ZnO is employed as the EIL and the other functional layers retain the same to the control device. According to the analysis in other literature [24], the exciton recombination zone should be located at QD/CBP interface. The optical properties of QDs used in our work are shown in Fig. 3. As can be seen from the absorption and PL spectra of the CdSe/CdS/ZnS QDs (in toluene) shown in Fig. 3(a), the PL peak is located at 598 nm and the full-width at half-maximum (FWHM) is 42.1 nm. The quantum yields of CdSe/CdS/ZnS QDs are respective 65% and 50% in solution and solid powder form measured with an integrating sphere. The decrease of quantum yield of the QDs relative to the solution may originate from the interactions (energy transfer and/or charge transfer) between QDs in the closed packed film. The TEM image of the CdSe/CdS/ZnS QDs is shown in Fig. 3(b) and the average diameter of the QDs is ~7.7 nm.

 

Fig. 3 (a) PL and Abs spectra of QDs in solution, and (b) TEM image of QDs.

Download Full Size | PDF

Figure 4(a) shows the voltage-current density curves of both QLEDs. We can see that the QLED based on s-ZnO EIL possesses much lower current density at low applied voltage range less than 3 V. Rather, the s-ZnO based device shows smaller leakage current than that of n-ZnO containing device, which might be attributed to the much rougher surface of n-ZnO as shown in Fig. 2(a). Larger current at high bias (larger than 6 V) is achieved for the n-based device, which should be due to the better electron mobility properties of n-ZnO because of the higher crystalline quality of ZnO nanoparticles than that of the amorphous s-ZnO film. Therefore, the maximum luminance of n-ZnO based device is higher than that of s-ZnO based device due to the better electron injection capacity of n-ZnO as shown in Fig. 4(b). Figure 4(c) shows the voltage-current efficiency properties of QLEDs. It is worth noting that the s-ZnO based device possesses higher efficiency than that of n-ZnO containing QLED at low voltage, which is mainly due to the lower leakage current.

 

Fig. 4 Photo-electrical properties of QLEDs with n-ZnO and s-ZnO as EIL. (a) voltage-current density, (b) voltage-luminance, (c) voltage-current efficiency, and (d) EL spectra.

Download Full Size | PDF

Figure 4(d) shows the EL spectra of both QLEDs. We can see that the EL spectra are nearly the same for these two devices, which indicates that the excitons are always formed in QDs despite different ZnO EILs used in the QLEDs. In addition, the EL spectra are almost consistent with the PL one measured in the solution. In detail, the FWHM of EL spectra are also 45.9 nm, which is similar to that of the PL one (42.1 nm). In addition, we can see that the EL spectra show a little red shift of 6 nm relative to that of the PL one in toluene. This should be due to the energy transfer from small size QDs (with a shorter emission wavelength) to the large ones (with a longer emission wavelength). This result further match up with the decrease of quantum yield when the QDs are formed a film from solution. However, this well complete PL-EL spectral overlap demonstrates the efficient recombination of holes and electrons on the QDs in our QLEDs.

To further clarify the effects of s-ZnO layer on the charge injection, the interfacial energy level is determined by employing the UPS measurement [25]. Figure 5 shows the UPS spectra of the ITO, ITO/s-ZnO films. There are clear spectra changes after the spin coating of s-ZnO film compared to that of neat ITO, and both the highest occupied molecular orbital (HOMO) onset position and the cut off position shift towards the higher binding energy, respectively. By analysis the UPS data [26], we can calculate that the work function of ITO/s-ZnO is about 3.9 eV, 0.8 eV lower than that of bare ITO, which will results in the lower electron injection barrier and benefit to get a good device performance, demonstrating that s-ZnO is a good EIL in QLEDs.

 

Fig. 5 UPS spectrum of bare ITO and s-ZnO/ITO samples.

Download Full Size | PDF

4. Conclusion

In conclusion, the s-ZnO is demonstrated as the EIL for efficient QLEDs and the device performance is comparable to that of common used n-ZnO EIL. AFM data show that the surface of s-ZnO film is smoother than that of n-ZnO film because the s-ZnO obtained by spin-coating sol-gel precursor solution is a continuous amorphous oxide film. In contrary, the n-ZnO layer deposited from ZnO nanoparticle solution is a close packed ZnO nanoparticle film, which will result in a larger RMS. Our work offers a reliable and cost-effective ZnO EIL to fabricate QLEDs. We believe that there are no fundamental obstacles to fabricate differently colored (green and blue) QLEDs with this s-ZnO EIL. These results provide a better alternative for inorganic oxide EIL and offer a practicable platform for display and lighting applications based on QLEDs as well.