1 Introduction

The organic light-emitting diodes (OLEDs) more often have been investigated in both mercantile and scientific fields, because of their exclusive advantages such as light weight, slender structure, high resolution and low driving voltage [1]. Recently, polymer light-emitting devices (PLEDs) attracted a great attention due to their advantages of emission in a wide visible region, solution processing, including screen printing and inkjet deposition, large area coverage, low and poor consumption in application as light sources for optical signal circuits, etc. [2]. Although important progress has been achieved on the presentation of PLEDs, they still need additional improvements such as stability, brightness and lifetime to attract commercial applications [3]. The basic structure of a PLED consists of several organic thin films sandwiched between two electrodes. Nowadays, due to their specific properties, the application of nanoparticles and nanostructures is attracting a lot of research efforts, to enhance the performance of PLEDs. Several studies showed that doped NPs enhanced the abilities of charge injection and charge transport and consequently the turn-on voltage of the device reduced. Therefore, the device characteristics in current density were obviously improved by doping nanoparticles (NPs) into the hole transport (HTL) layer of PLEDs [1, 4, 5]. It has been demonstrated that the optoelectronic efficiency of OLEDs can be practically improved by measuring doped nanoparticles or nanostructures in and/or between organic thin films [6]. By doping suitable inorganic nanoparticles, the electrical properties of PEDOT:PSS can be improved, and the ensuing of PEDOT:PSS nanocomposite layers in PLEDs makes the injection of holes easy [7, 8]. In this study, we have made an organometallic complex based on aluminum ions, then the effect of PEDOT:PSS/ZnO nanocomposite hole–injection layer on the enhancement of performance of PLEDs has been investigated. Simulation device was also done by constructing the model and choosing appropriate parameters. The experimental data were collected and the results interpreted both qualitatively and quantitatively. Also experimental results and simulated data were compared.

2 Experimental

2.1 Materials

The poly(9-vinylcarbazole) (PVK), 2-(4-biphenylyl)-5-phenyl-1,3,4-oxadiazole (PBD), poly(3,4-ethylenedioxythiophene) polystyrene sulfonate and ZnO NPs were purchased from Aldrich.

2.2 Synthesis

8-Hydroxyquinoline (Aldrich, 4.35 g) was dissolved in a mixture of 15 mL methanol (Merck); it was stirred well until orange transparent solution was obtained. Separately, Al (NO3)3·9H2O (99%, Aldrich, 3.75 g) was dissolved in methanol and stirred well till crystal transparent solution was obtained. These two mixtures were added together, and the resulted suspension was stirred for 20 min. The final product was dried in the room temperature and was used for further experiments.

2.3 Physical measurements

The crystal structures of Al complex were determined by means of X-ray diffraction (XRD, Philip X’pert X-ray diffractometer with Cu Kα radiation, λ = 1.5418 Å). Infrared spectra from 200 to 4000 cm−1 were recorded on a Shimadzu 470 FT-IR instrument, using KBr pellets. Voltammetric experiments were performed using a m-Autolab Type III electrochemical system. A conventional three-electrode cell consisting of a glassy carbon working electrode, a platinum wire counter electrode and a saturated Ag/AgCl reference electrode was used for voltammetric experiments. The current–voltage characteristics of the devices were measured by Keithley, as 2400 source meter. The EL and PL spectra were measured with USB2000 and HR4000 ocean optic, under ambient conditions. All measurements were carried out immediately after fabricating the devices.

3 Fabrication OLED

3.1 Simulation method

Simulation is an influential method to examine the underlying mechanism of the devices. More importantly, it can tell us to what extent the mechanism can affect the device performance [9]. Meanwhile, simulation can regain detailed information, such as the charge carrier distribution which is difficult to be directly measured through experiments. The first step in the simulation of PLED devices is to divide the organic layers into a set of discrete monolayers, sandwiched between an anode and a cathode [9–11]. The simulations presented in this work are based on injection, transport and recombination of charge carriers, as well as the electric field distribution. Injection of charge carrier from a metal electrode into an inorganic semiconductor can be described by the thermionic emission model [12], which assumes that the current flow across the barrier is dependent on the barrier height only [11, 13, 14]. Based on this model, the injection current can be calculated by the following equations:

$$J_{\text{th}} = AT^{2} \exp \left( { - \frac{{\phi_{b} B\sqrt E }}{kT}} \right)$$
$$A = \frac{{4\pi emk^{2} }}{{h^{3} }} \quad B = \sqrt {e^{3} /4\pi \varepsilon_{r} \varepsilon_{0} }$$

The energy and position disorder impedes the transport and causes a low mobility of organic materials. Bässler [15] used the Monte Carlo method to analyze the transport behavior in disordered materials. The dependence of mobility on electric field can be described by the Poole–Frenkel equation [15, 16] as follows

$$\mu \left( E \right) = \mu_{0}^{*} \exp \left( {\gamma^{*} \sqrt E } \right)$$

The electric field of the device is calculated by the following equation [9, 10, 17–20]:

$$\frac{\partial E}{\partial x} = \frac{{e\left( {p - n} \right)}}{{\varepsilon_{r} \varepsilon_{0} }}$$

3.2 Fabrication of OLED by utilization of prepared complex

3.2.1 The first step of the fabrication process

ITO substrates were patterned by the etching process using an acidic mixture of HCl and HNO3 as etching agent. Patterned ITO glasses were cleaned with acetone, dichloromethane, methanol and deionized water under an ultrasonic bath for 10 min and then dried in an oven for 20 min. The PEDOT:PSS and PEDOT:PSS:ZnO layers were spin-coated on ITO glasses. To remove residual water, the layers were baked into the oven at 120 °C for 20 min. The PVK in dichloromethane solution was spin-coated on the PEDOT:PSS and PEDOT:PSS:ZnO layers. The Alq3 and PBD were spin-coated in separated dichloromethane solution and baked in the oven, respectively, as the former stage. The devices were prepared after thermally evaporating Al that is as a cathode layer. The evaporation rate of Al was kept at 0.6–1 Å/s.

3.2.2 Fabrication of OLED

The steps of the fabrication process of a PLED device are schematically demonstrated in Fig. 1. In this study, four OLED devices were manufactured with different organic layer thickness. Active layers were used with an indium tin oxide (ITO)/PEDOT:PSS (35 nm) or PEDOT:PSS:ZnO (35 nm)/PVK (70 nm)/Alq3 (0.0032, 0.0038 M) (60 nm)/PBD(70 nm)/Al (100 nm) device structure. The size of ZnO NPs was about 31 nm. OLED devices were fabricated with the following configurations (Table 1).

Fig. 1
figure 1

Schematics illustrating the fabrication steps of PLED device

Full size image
Table 1 Device structures
Full size table

ITO/PEDOT:PSS/PVK/Alq3 (0.0032 M)/PBD/Al (device 1), ITO/PEDOT:PSS/PVK/Alq3 (0.0038 M)/PBD/Al (device 2), ITO/PEDOT:PSS:ZnO/PVK/Alq3 (0.0032 M)/PBD/Al (device 3) and ITO/PEDOT:PSS:ZnO/PVK/Alq3(0.0038 M)/PBD/Al (device 4).

4 Results and discussion

4.1 Characterization synthesis

Crystal structures of Alq3 were determined by X-ray diffraction (Fig. 2). The diffraction peaks can be indexed to tris-(8-hydroxyquinoline) aluminum (Alq3) (JCPDS 26-1550). The determined α-phase of XRD pattern of the meridional isomer Alq3 is reported by Michael Cole et al. [21]. No characteristic peaks of impurity were detected; therefore, the products are pure Alq3. This result almost matches with the results observed from the FT-IR curve. The spectra are shown in the range of 200–4000 cm−1 in Fig. 3. The characteristic peaks of Alq3, which are the bands centered at 600–800 cm−1, come from the vibrations of quinoline, and the band at 400–600 cm−1 can be attributed to the stretching vibration of Al ion with ligands. As examples of them, we have C–C stretching vibrations (1502, 1469 cm−1), C–O stretching vibrations (1035, 1111, 1469 cm−1), aromatic amine (1300 cm−1), aromatic stretching C=C (1469 cm−1), Al–O–Al (758 cm−1), Al–O (557, 644 cm−1) and Al–N stretching vibrations (416, 460 cm−1). The peak at ~3416 cm−1 indicates the amount of moisture (H2O) present in the sample [22–28]. Figure 4 shows the absorption and emission spectra of Alq3 in aqueous solution. The complex Alq3 exhibits an ultraviolet absorption peak in the range of 300–450 nm, and the maximum absorption peaks are located at 292 (π–π *) and 374 (n–π *) nm, respectively [29–33]. The band at 374 nm has multiple electronic origins, that is, this band is a superposition of two or more electronic transitions, for more information see Burrows et al. [34]. As it is shown in Fig. 4, the emission peak of Alq3 in aqueous solution is 524.39 nm, and we can clearly observe the green light.

Fig. 2
figure 2

XRD pattern of yellowish green Alq3 powder

Full size image
Fig. 3
figure 3

FT-IR absorption spectrum of Alq3

Full size image
Fig. 4
figure 4

UV–Vis absorption and emission (norm) spectrum of Alq3 aqueous solution

Full size image

4.2 Energy levels of prepared complex

One of the most accurate methods to characterize the organic materials and estimation about energy band diagram is cyclic voltammetry (CV). The energy band diagram of the Alq3 complex was determined from the HOMO and LUMO energy levels and the band gaps was calculated from cyclic voltammetry [35, 36]. The highest occupied molecular orbital (HOMO) energy level of the complexes was evaluated by cyclic voltammetry on an electrochemical work station. The lowest unoccupied molecular orbital (LUMO) energy level was appraised by subtracting the energy gap from the energy of HOMO, and the energy gap was determined by the absorption edge of the UV absorption spectrum of the complex in the solution. Based on cyclic voltammetry results, Alq3 shows E ox, which is equal to E HOMO from the equation E HOMO = [(E ox − E 1/2(ferrocene)) + 4.8] eV. Figure 5 shows cyclic voltammetry of aluminum complex with the scan rate of 100 mV/s. The calculated E HOMO, E LUMO and E g values are listed in Table 2. The optical band gap of the complex Al (0.0032, 0.0038 M) was determined from the absorption onset (Fig. 6). The corrected base line of the absorption spectra was drawn horizontally. In order to determine the absorption edge, a straight line to the right-hand side of the maximum peak was fitted, and then, its intersection with the corrected baseline was determined [37]. The optical band gap we obtained for Alq3 corresponds well to the value reported [38].

Fig. 5
figure 5

Cyclic voltammograms of aluminum complex at scan rate of 100 mV/s

Full size image
Table 2 Physical properties of aluminum complexes
Full size table
Fig. 6
figure 6

Determination of the band gap by using the absorption onset

Full size image

4.3 OLED device

Figure 7 shows the XRD patterns of layers with and without ZnO NPs. As shown in Fig. 7, the intensity and width of the peak decreased when the ZnO NPs were doped with PEDOT:PSS. XRD spectrum parameters for layers with and without ZnO NPs were calculated and are reported in Table 3. According to the Scherrer equation [39]:

$$L = \frac{0.9\lambda }{\beta \cos \theta }$$
Fig. 7
figure 7

XRD pattern of layers with and without ZnO nanoparticles

Full size image
Table 3 XRD spectrum parameters for layers with and without ZnO nanoparticles
Full size table

It can be concluded from Table 3 that the crystal size for layers with ZnO NPs increased and it is followed by the reduction of band gap. Consequently, there have been more effective electron injection/transfer. Figure 8 shows the EL spectra that are normalized at the main emissive peak at ~515 nm. The device without ZnO NPs shows the typical EL spectrum where the emission peak at ~515 nm is larger than the vibrancy peak at longer wavelengths. In comparison, the long-wavelength EL peak increases relatively, with the EL peaks slightly shifted to longer wavelengths when ZnO NPs were doped with PEDOT:PSS. This may be because of the optical interference effects. Interference effects depend on the amount of ZnO NPs or changes of the spectral characteristics of the EL emission. All electrical properties are shown in Figs. 9 and 10. Figure 9 shows the current density–voltage (J–V) as well as characteristics of prepared devices. It is clear from the Fig. 9 that the current density of device increases significantly by doped ZnO NPs compared to the device without the ZnO NPs: the turn-on voltage decreases by about 3 V and the maximum luminance increases more than ten times when doped ZnO NPs were used. The doping of ZnO can effectively enhance the conductivity in the hole–injection layer and reduce the turn-on voltage that made the progress of hole injection more smoothly and led to a higher overall device efficiency. To confirm the outcome of the simulations, the experimental J–V spectra of the devices were compared with the resulting spectra of simulation. Figure 10 shows experimental and simulated current density–voltage (J–V) spectra of four OLED structures: ITO/PEDOT:PSS/PVK/Alq3 (0.0032 M)/PBD/Al (device 1), ITO/PEDOT:PSS/PVK/Alq3 (0.0038 M)/PBD/Al (device 2), ITO/PEDOT:PSS:ZnO/PVK/Alq3 (0.0032 M)/PBD/Al (device 3) and ITO/PEDOT:PSS:ZnO/PVK/Alq3 (0.0038 M)/PBD/Al (device 4). As it is shown in the Fig. 10, good agreement was observed between simulated and experimental curves. The difference between the curves can be attributed to the carrier density.

Fig. 8
figure 8

Electroluminescence spectra of four different structure PLEDs

Full size image
Fig. 9
figure 9

Current density versus voltage characteristic of four samples with and without ZnO nanoparticles

Full size image
Fig. 10
figure 10

Experimental and simulated current density–voltage (J–V) spectra of four OLED structures

Full size image

5 Conclusion

In this study, an Al complex was synthesized and it is demonstrated that light-emitting layers of Alq3 could be formed via spin coating. The properties of Alq3 particles were studied by XRD, FT-IR, UV–Vis, CV and PL analyses. FT-IR results confirm the completion of quinoline ring formation and presence of quinoline structure in the metal complex. In the absorption spectra, the complexes show two absorption bands at 292 and 374 nm. The emission peak of Alq3 in aqueous solution is 524.39 nm, and the green light was observed clearly. The XRD and PL studies showed that the prepared yellowish green Alq3 powder in this experiment is an α-phase, that is, one of the four phases mentioned above. Examination of a device operation with an Alq3 EML confirmed that the electron transport and the green emission characteristics of Alq3 were retainable by Alq3, but the EL efficiency was very low because of the deficiency of holes that recombine radiatively with electrons. We investigated the effects of ZnO NPs doped with PEDOT:PSS on the performance of OLEDs. Then electron transport (ETL), hole transport (HTL) and emissive layers were spin-coated with spin speeds of 2000, 2000 and 1500 rpm, respectively. As a result, experiments showed that the electrical and optical efficiency in samples with ZnO NPs are appreciably higher than in samples without ZnO NPs. In summary, when ZnO NPs were used, the electroluminescence and J–V curves from PLEDs at voltages below the band gap voltage were observed. It can be concluded that the crystal size for layers with ZnO NPs increased and it is followed by the reduction of band gap. Consequently, there was more effective electron injection/transfer in device. The doping of ZnO can effectively enhance the conductivity in the hole–injection layer and reduce the turn-on voltage that made the progress of hole injection more smoothly and led to a higher overall device efficiency. In this study, the electrical properties of the highly efficient material systems for OLEDs were modeled successfully. Simulations of electron and hole-only devices verified the validity and accuracy of the used models and parameters. Also good agreement between simulated and experimental curves was observed.