1. Introduction

Much research interest has developed recently with regard to solution-processable organic (or polymeric) semiconducting devices, including organic light-emitting devices (OLEDs) and organic photovoltaic (PV) devices (OPVs) [1–16 ]. Particularly, the focus has been on the development of new semiconducting materials and device structures with the aim of realizing cost-effective, low-temperature devices that are fast and simple to fabricate, and that offer devices with as large an area as possible [1–14 ]. To realize these goals, researchers have mainly focused on improving device efficiency and stability levels and on enhancing the simplicity of the production process. In recent years, the electroluminescent (EL) performance of simple solution-processed OLEDs and OPVs has been enhanced considerably. For example, by introducing phosphorescent light-emitting dopants into the emission layers (EMLs), some solution-processed OLEDs have shown luminous efficiency (LE) of nearly 100 lm/W [15]. Moreover, a body of important research now exists on coating processes such as roll-to-roll coating techniques, including the use of new device structures and functional compounds [7–10 ]. To date, however, the performance capabilities of solution-processed devices, especially for OLEDs, do not match those of devices fabricated using conventional complicated vacuum-deposition techniques. Thus, there is a strong need to improve the device performance of solution-processed devices further given that there is still some way to go before they can compete on equal terms with vacuum-deposited devices.

One way to improve the efficiency and stability of solution-processable devices is by reducing the height of the potential barrier at the electrode contacts with the organic/polymeric active layers in an effort to make the injection or extraction of the charge carriers as efficient as possible. While the barrier height at the cathode contacts can be modified by tuning the work function of the metal cathode to the lowest unoccupied molecular orbital (LUMO) of the active layer [17–19 ], the barrier height at anode contacts can be optimized by introducing a hole-selective layer (HSL), such as the hole-injection layer (HIL) in OLEDs or the hole-collecting layers (HCLs) in OPVs, on top of an anode, e.g., a transparent indium tin oxide (ITO) anode. The HSL must have an energy level that matches as closely as possible the energy level that lies between the highest occupied molecular orbital (HOMO) of the active layer and the work function of ITO anode, because the work function (4.5-5.0 eV) of ITO is lower than the typical HOMO of the active layer [17–19 ]. Several researchers have attempted to match the energy band using various material systems and to obtain flat and uniform layers. Among them, a p-type conducting polymeric layer of poly(3,4-ethylenedioxythiophene): poly(styrenesulfonate) (PEDOT:PSS) has often been used as the HSL. However, although the PEDOT:PSS HSL has successfully been applied to match the energy level of ITO, PEDOT:PSS is highly acidic due to its sulfate ions (pH ~1.5), which can damage the surface of the ITO and dissolve the ions that may migrate into the stacked active layer, gradually reducing device performance levels over time [20,21 ]. This problem makes it necessary to consider other neutral materials for HSLs which are also chemically stable, mechanically uniform, and solution-processable in order to minimize any undesired effects in solution-processable devices. Other alternatives to PEDOT:PSS have been tested, specifically transition-metal oxides such as MoO₃, V₂O₅, NiO, and WO₃ based on their favorable energy level alignments, using various different deposition methods [22–25 ]. For example, polymer- and/or small-molecule-based OPVs with sol-gel-processed MoO₃ film exhibit device performance levels comparable to those with PEDOT:PSS layers [22]. However, such transition-metal-oxide HSLs still produce device performances that are only comparable with those of the PEDOT:PSS layer.

Recently, several important attempts have been made to use solution-processable buffer layers of neutral graphene oxide (GO) sheets as a HSL on ITO anodes [26–29 ]. Aqueous-processable GO, which is a two-dimensional sheet of graphene functionalized with an oxygen group, has a unique electronic structure due to its mixed sp² and sp³ hybridizations [30]. This heterogeneous property, together with its work function and large bandgap, make it a potential HSL candidate in organic devices, with none of the problems associated with ITO erosion [17]. Hole-selective GO films have been tested in OPVs with bulk heterojunction (BHJ) OPV layers of poly(3-hexylthiophene) (P3HT) and fullerene-derivative phenyl-C₆₁-butyric acid methyl ester (PCBM) blends, providing a power conversion efficiency of 3.25% [26]. It has also been demonstrated that a bilayer of GO and PEDOT:PSS (i.e., double-stacked GO/PEDOT:PSS) could protect the ITO from ion diffusion and acid etching. Thus, stacked double HSLs of GO/PEDOT:PSS may lead to improved device performance capabilities for both OLEDs and OPVs which are somewhat comparable with (or possibly slightly inferior to) those of reference devices with a conventional single PEDOT:PSS layer [27,28 ]. Thus, the further development of solution-processable HSL(s) for improved and balanced hole-injection/electron-blocking properties is clearly needed, as is a novel device structure that affords high performance and efficiency. Quite recently, we reported the use of thin and homogeneous triple-stacked layers of GO, MoO₃, and PEDOT:PSS films (i.e., GO/MoO₃/PEDOT:PSS) [29] fabricated by means of horizontal-dip (H-dip) coating [13,28,29 ] on ITO substrates in an effort to replace the conventional single PEDOT:PSS layer or double-stacked GO/PEDOT:PSS layers. By employing triple-stacked layers of GO/MoO₃/PEDOT:PSS, we showed highly improved device performances in solution-processed devices and demonstrated the feasibility of fabricating large-area solution-processable OLEDs and OPVs. These improved properties would make the GO/MoO3/PEDOT:PSS layers superior to PEDOT:PSS but without the aforementioned drawbacks, allowing them to be used in solution-processable OLEDs, OPVs, and other organic electronic devices.

We herein describe our attempts to understand the impedance spectroscopy (IS) characteristics of solution-processable OLEDs including the multi-stacked HSLs of GO, MoO₃ and PEDOT:PSS used as HILs. Thin and homogeneous stacked HSLs consisting of GO, MoO₃, and PEDOT:PSS films were fabricated using a H-dip-coating method [13,28,29 ] on ITO substrates. We report on the influence of the H-dip-coated stacked HILs on the IS behaviors of solution-processed OLEDs. By employing triple-stacked layers of GO/MoO₃/PEDOT:PSS, we confirmed highly improved device performance and obtained characteristic IS properties of solution-processed OLEDs. We attribute this device performance to the all-water-processable triple-stacked HILs of the GO/MoO₃/PEDOT:PSS used, leading to good film-forming capabilities and providing not only good matching of the energy levels between adjacent layers for efficient hole-injection/electron blocking but also improved impedance characteristics in the form of a reduced level of parallel resistance. Moreover, we studied the relaxation frequency of an OLED with triple GO/MoO₃/PEDOT:PSS HILs to understand the carrier (hole) injection process in the OLEDs.

2. Experiment

In this study, the solution-processed semiconducting layers used were fabricated on glass substrates pre-coated with an ITO layer (80 nm, 30 Ω/square) after routine cleaning of the ITO substrate. As shown in Fig. 1(a) , using a simple H-dip-coating method [13,28,29 ], HILs were prepared in the form of triple-stacked layers of GO, MoO₃, and PEDOT:PSS. A GO layer was H-dip-coated onto the ITO anode using an aqueous GO dispersion [28,31 ] and was then annealed at 150°C. A MoO₃ layer was also H-dip-coated using an aqueous MoO₃ solution consisting of ammonium heptamolybdate ((NH₄)₆Mo₇O₂₄·4H₂O, Sigma-Aldrich) as a precursor [22] and was subsequently annealed at 150°C for conversion to MoO₃. A layer of PEDOT:PSS (Clevios PVP AI 4083, H.C. Starck, 1%) was also H-dip-coated and then baked at 120°C.

 

Fig. 1 (a) Left: Photograph and schematic illustration of the H-dip-coating process with the gap height h₀ and coating speed U. Right: Chemical structures of graphene oxide (GO) and ammonium heptamolybdate (a MoO₃ precursor). (b) 10 × 5 μm² AFM topography images of the H-dip-coated MoO₃ (left upper), GO (left lower), GO/MoO₃ (right upper), and GO/MoO₃/PEDOT:PSS (right lower) layers on flat substrates. (c) Current density-voltage (J-V, Left) and luminance-voltage (L-V, Right) characteristics of representative OLEDs studied here.

Download Full Size | PDF

To form an EML, a blended solution was spin-coated onto the ITO layer, which was pre-coated with the HIL(s). For the blended solution, we used hole-transporting N,N'-diphenyl-N,N'-bis(3-methylphenyl)-1,1'biphenyl-4,4'-diamine (TPD, Sigma-Aldrich), electron-transporting 2-(4-biphenylyl)-5-(4-tert-butylphenyl)-1,3,4-oxadiazole (Bu-PBD, Sigma-Aldrich), green-emitting tris(2-phenylpyridinato) iridium (Ir(ppy)₃, Lumtec), and poly(vinylcarbazole) (PVK, Sigma-Aldrich) in a mixed solvent of 1,2-dichloroethane and chloroform [13,28,29 ]. Then, an electron-injecting Cs₂CO₃ layer (2 nm, Sigma-Aldrich) and an Al cathode (50 nm) were evaporated on the EML (85 nm) via thermal deposition. Therefore, the investigated OLED configuration was an ITO anode/H-dip-coated HIL(s)/EML/Cs₂CO₃ electron-injecting layer/Al cathode. More detailed information related to the fabrication of functional layers and devices is available in the literature [29].

In order to investigate the film quality of the fabricated HILs, the surface roughness of the HIL(s) was observed using atomic force microscopy (AFM, Nanosurf easyScan 2 AFM, Nanosurf AG Switzerland Inc.). A Chroma Meter CS-200 device (Konica Minolta Sensing, Inc.) and a source meter (Keithley 2400) were used to measure the current density-voltage-luminance (J-V-L) characteristics of the fabricated OLEDs. IS measurements of the fabricated devices were made using an HP 4192A. In the IS measurements, the AC oscillating amplitude was set to be as low as 100 mV (root mean square, RMS) to maintain the linearity of the response. The IS measurements were performed in the dark for different bias voltages. Characterization of the device was carried out at room temperature under ambient conditions, without encapsulation.

3. Results and discussion

Using the simple H-dip-coating method [13,28,29 ], we prepared functional layers of GO, MoO₃, and PEDOT:PSS, as shown in Fig. 1(a). The H-dip-coating technique used here is a self-metered coating method for the deposition of a solution in a controlled fashion. The thickness of the coated layer can be described by the associated drag-out problem proposed by Landau and Levich [13,28,29,32 ]. As a typical example, we tested the control of the thickness of the H-dip-coated MoO₃ layer as a function of the coating speed U; the thickness of the H-dip-coated MoO₃ layer showed an increase from ca. 5 nm to ca. 20 nm with an increase in U from 0.3 to 2.0 cm/s, confirming that H-dip-coating allows the precise nanoscale thickness control of an MoO₃ film. It was also noted that the optical properties of 7.3-nm-thick MoO₃ films coated onto ITO glass substrates had a high transmittance minima of ca. 76% at a wavelength of about 400 nm, similar to those of a 4.5-nm-thick GO-coated ITO substrate and a bare ITO substrate. More detailed information pertaining to such optical results is available in the literature [29].

Next, we determined the morphologies of the fabricated layers by observing the surface roughness of the layers on flat substrates (RMS roughness ~2.5 nm) by means of AFM. Figure 1(b) shows the AFM morphologies of 10 × 5 µm² scanned areas of H-dip-coated MoO₃ layers deposited onto flat substrates [left upper panel] and onto GO pre-coated substrates [right upper panel], in comparison with those of a GO layer [left lower panel] and GO/MoO₃/PEDOT:PSS layers deposited onto flat substrates [right lower panel]. As shown in the figure, H-dip-coating allows the functional materials of GO and MoO₃ to form films of a high quality. The AFM investigation clearly reveals that the coated films are fairly uniform and smooth; the RMS surface roughness levels of GO, MoO₃, GO/MoO₃, and GO/MoO₃/PEDOT:PSS are only ca. 0.80, 0.63, 0.99, and 0.92 nm, respectively, showing good film coverage in each case. Furthermore, the surface roughness was nearly identical at different positions in the investigated layers. Such smoothness and uniformity of the H-dip-coated layers clearly indicate that the HSLs produced by H-dip-coating may be suitable for the fabrication of solution-processed OLEDs. It is also important to note that after thermal annealing, the water-processed GO and MoO₃ layers were not destroyed by the subsequent H-dip-coating step used to fabricate the multi-stacked layers, in contrast to the more detrimental effects on the PEDOT:PSS layer.

Next, we investigated the device performance capabilities and IS characteristics of green OLEDs including HILs of H-dip-coated GO, MoO₃, and PEDOT:PSS in an effort to understand the hole-injecting properties of the HILs studied. For comparison, we also fabricated other OLEDs using different types of HILs; three representative examples used were as follows: an OLED with H-dip-coated triple-stacked HILs of GO/MoO₃/PEDOT:PSS (sample OLED), an OLED with an H-dip-coated single HIL of PEDOT:PSS (reference OLED 1), and an OLED with H-dip-coated double-stacked HILs of GO/PEDOT:PSS (reference OLED 2). In the OLEDs studied, the functional layers in the HIL(s) were deposited using the H-dip-coating process at fixed thicknesses of GO of 4.5 nm, MoO₃ of 7.3 nm, and PEDOT:PSS of 40.0 nm, at which OLED performance was found to be optimal.

The J-V-L characteristics of the three representative devices are also shown in Fig. 1(c). As shown in the figure, for all OLEDs tested, the slopes of the J-V curves [Left panel of Fig. 1(c)] represent excellent diode behaviors, indicating good coverage of the H-dip-coated functional layers, whereas the sample OLED (with triple-stacked HILs of GO/MoO₃/PEDOT:PSS) showed apparently excellent current flow characteristics with a sharper increase in the J-V curve than reference OLEDs 1 and 2. These J-V curves clearly show the considerably better hole-injecting characteristics of the triple-stacked GO/MoO₃/PEDOT:PSS HILs compared with the other reference OLEDs, especially with reference to the single PEDOT:PSS HIL used in reference OLED 1.

Regarding the J-V curves, it is also clear from the L-V characteristics [Right panel of Fig. 1(c)] that the OLEDs showed different EL characteristics and that the sample OLEDs outperformed the others, indicating the significantly reduced hole-injection barrier due to the introduction of the multi-stacked HILs. It was also noted that the current efficiencies were significantly improved for the sample device with the triple-stacked GO/MoO₃/PEDOT:PSS HILs. These results clearly indicate that the triple-stacked GO/MoO₃/PEDOT:PSS HILs in the sample OLED provide the best outcomes among all of the HILs studied. This improved device performance of the sample OLED is attributed to the excellent hole-injecting and/or electron-blocking properties and to the better matching of energy levels of the GO/MoO₃/PEDOT:PSS HILs used between the ITO anode and the EML as compared to those of the other schemes. More detailed information pertaining to the EL performance of the OLEDs studied here is available in the literature [29]. Based on these observations, it is clear that the use of all-water-processable triple-stacked GO/MoO₃/PEDOT:PSS HILs results in greater overall brightness and efficiency of solution-processable OLEDs, as reported previously [29].

In order to understand the underlying mechanism behind the considerable improvement in the device performance capabilities of the sample OLEDs, we investigated the impedance characteristics of different devices using IS measurements, which can be used to characterize the electrical properties of materials and their interfaces [33,34 ]. In these IS observations, we assessed the effects of different types of HILs in OLEDs on the electrical characteristics. The IS characteristics of the representative devices are presented in Fig. 2 , which shows plots of the real (Re(Z)) and imaginary (-Im(Z)) parts of the complex impedance (Z) (Cole-Cole plots) for the three OLEDs at frequencies f between 20 Hz and 13 MHz at bias voltages of 4.0 V [Fig. 2(a)] and 5.0 V [Fig. 2(b)]. In these plots, the implicit variation is in the frequency, which increases from right to left. It is clear from the figures that the IS results for the investigated devices resemble those of a single-layer and single-carrier (hole-only) device, with a single semicircle in the Cole-Cole plots that decreases in size with an increase in the bias voltage [33,34 ]. Under low bias voltage conditions, the devices studied can be considered as hole-dominant devices, allowing us therefore to consider a single parallel resistor 𝑅_p and capacitor 𝐶_p network with series resistance 𝑅_s to be an equivalent circuit for the investigated OLEDs [inset in Fig. 2(b)] [35].

 

Fig. 2 Cole–Cole plots of the real and the imaginary parts of the complex impedance (Z) for the solution-processed OLEDs with different types of HILs at bias voltages of 4.0 V (a) and 5.0 V (b). The inset shows an equivalent circuit describing the impedance spectroscopy. The solid curves show the simulated results according to the equivalent circuit.

Download Full Size | PDF

Analyses of the equivalent circuit using the parameters 𝑅_s, 𝐶_p, and 𝑅_p [Figs. 3(a), 3(b), and 3(c) , respectively] for reference OLED 1 yield an estimated value for 𝑅_s of approximately 31.1 ohms and a value for the bulk capacitance 𝐶_p of about 2.60 nF. These values are nearly independent of the bias, as shown in the figure. However, the bulk resistance 𝑅_p decreases rapidly with an increase in the bias voltage. These findings clearly indicate that as the bias is increased, more charge carriers (holes) are injected into the device, resulting in a decrease in the dielectric relaxation time τ (τ = 𝐶_p𝑅_p) [36,37 ]. In the same manner, reference OLED 2 with the GO/PEDOT:PSS and the sample OLED with the GO/MoO₃/PEDOT:PSS HILs show 𝑅_s values of 31.3 and 30.3 ohms, respectively, and corresponding 𝐶_p values of 2.55 and 2.60 nF, respectively. These are each quite similar to the values obtained for reference OLED 1. In contrast, the 𝑅_p values for the sample OLED device with the triple GO/MoO₃/PEDOT:PSS HILs are much lower than those of reference OLEDs 1 and 2 [Fig. 3(c)].

 

Fig. 3 Variations of the obtained series resistance R_S (a), parallel capacitance C_P (b), and parallel resistance R_P (c) plotted against the bias voltage for solution-processed OLEDs with different types of HILs.

Download Full Size | PDF

Next, in order to obtain a better understanding of the current flows in the OLEDs studied, we investigated the dependence of 𝑅_p on the voltage V. Figure 4(a) presents plots of log(𝑅_p) versus log(V) of the devices, showing that log(𝑅_p) is linearly dependent on log(V). Based on the space-charge-limited current (SCLC) theory, the current J is given by J ∝ V ^m, which provides information about the exponent m, i.e., the slope of log(𝑅_p) versus log(V) [38–40 ]. The evaluated 𝑚 value is ca. 2.3 for the sample OLED in the applied voltage range of 3 V < V < 6 V. This 𝑚 value implies that the conductance of the major charge (hole) carriers for the sample OLEDs follows SCLC theory, with trap-free space-charge-limited conductivity (m = 2), in contrast to the exponential trap distributions (m > 2) for reference OLED 1 (m = 4.9) and reference OLED 2 (m = 2.8) [38–40 ]. These IS results clearly indicate that the hole carrier injection in the sample OLED with the triple GO/MoO₃/PEDOT:PSS HILs is more efficient [36,37 ] than that in any of the other devices tested.

 

Fig. 4 Variations of the obtained log (R_P) vs. log (V) (a) and relaxation frequency f_R plotted against the bias voltage V (b) for solution-processed OLEDs with different types of HILs.

Download Full Size | PDF

Moreover, the relaxation frequency f_R values estimated using the relationship of f_R = 1/(2πτ) [36,37 ] are much higher for the sample OLED with the triple GO/MoO₃/PEDOT:PSS HILs than for the other reference devices [Fig. 4(b)]. For example, at a bias voltage of 4.0 V, the sample OLED shows an f_R value of approximately 6.98 kHz, while the f_R value for the device is approximately 15.02 kHz at a bias voltage of 6.0 V. In contrast, at a bias voltage of 4.0 V, reference OLEDs 1 and 2 show f_R values of approximately 0.50 kHz and 2.55 kHz, respectively, while the corresponding f_R values of the devices are approximately 2.76 kHz and 7.81 kHz at a bias voltage of 6.0 V, clearly indicating that f_R for the sample OLED is greater than that of the reference devices. These IS results also clearly indicate that carrier (hole) injection in the sample OLED with the triple GO/MoO₃/PEDOT:PSS HILs is more efficient (with a higher relaxation frequency) [36,37 ] than that in any of the other devices.

Considering the above-mentioned IS results with the device performance data of the sample OLEDs, it is clear that the OLEDs with the H-dip-coated triple-stacked HILs of GO/MoO₃/PEDOT:PSS possess efficient hole-selective properties, i.e., hole carrier-injecting/electron-blocking and hole-extracting properties, superior to those of a conventional single PEDOT:PSS layer. Therefore, the use of all-water-processable triple-stacked GO/MoO₃/PEDOT:PSS HILs in solution-processable OLEDs is a sensible option as an alternative to the conventional single PEDOT:PSS layer due to their low resistance, high relaxation frequency (good charge injection/collection), high EL performance capabilities, and high efficiency, although multiple coating processes with the additional use of new materials as described here may increase the fabrication cost and become an important factor in the mass production of such devices. Moreover, the H-dip-coated GO layer in the triple-stacked HILs may increase the stability of devices significantly [29]. Therefore, the deposition of multi-stacked HSLs by H-dip-coating can also be applied to various organic electronic devices for high-throughput manufacturing processes such as roll-to-roll production because it is easier to use to realize the desired compositions of multi-stacked functional layers, in contrast to the more conventional complicated vacuum evaporation technique.

4. Summary

In summary, we have reported the IS characteristics of solution-processed OLEDs including all-water-processable triple-stacked hole-selective layers of GO/MoO₃/PEDOT:PSS. Using thin, smooth, and homogeneous GO/MoO₃/PEDOT:PSS deposited by H-dip-coating, we have shown that the device performances and IS behaviors of OLEDs with triple-stacked hole-selective layers were much better than those with a conventional single PEDOT:PSS layer due to their good hole-injecting/electron-blocking properties and better matching of the energy level with the EMLs with low bulk (parallel) resistance. These impedance behaviors clearly indicate that the conductance of the major charge (hole) carriers for the sample device with the H-dip-coated GO/MoO₃/PEDOT:PSS HILs follows SCLC theory with trap-free space-charge-limited conductivity, in contrast to the exponential trap distributions for the reference device with the single PEDOT:PSS HIL. Moreover, it was shown that the relaxation frequency of the sample device is much higher than that for the other reference devices. These results therefore indicate that carrier (hole) injection in the sample OLED is more efficient than that in any of the other devices, demonstrating that the introduction of thin stacked GO/MoO₃/PEDOT:PSS onto ITO substrates provides a solid foundation for high-performance solution-processable OLEDs. Therefore, these advanced triple-stacked hole-selective layers in solution-processable OLEDs can be applied in high-throughput manufacturing methods such as roll-to-roll production and can be used in various optoelectronic applications, such as lighting, display, and imaging applications.