Extremely high external quantum efficiency of inverted organic light-emitting diodes with low operation voltage and reduced efficiency roll-off by using sulfide-based double electron injection layers

Abstract

Inverted organic light-emitting diodes (IOLEDs) have great potential application in flat-panel displays. High energy consumption, efficiency roll-off, and poor electron injection are key issues limiting the use of IOLEDs. Here, we present IOLEDs with extremely low driving voltage, high efficiency and efficiency roll-up by employing double electron injection layers (D-EILs) composed of metal sulfide and cesium carbonate (Cs₂CO₃)-doped 4,7-diphenyl-1,10-phenanthroline (Bphen). We demonstrate that the use of D-EILs with metal sulfides can significantly improve the performance of IOLEDs. For a blue florescent device based on (2 nm-zinc sulfide)/Bphen: Cs₂CO₃, we achieve a power efficiency of 10.9 lm W⁻¹ at a luminance of 1000 cd m⁻², giving a turn-on voltage of 2.8 V. Notably, the external quantum efficiency increases from 6.9 to 7.5% and the current efficiency increases from 14.3 to 15.4 cd A⁻¹ with the rise in luminance from 1000 to 10 000 cd m⁻². Also, the copper sulfide-based device exhibits very-low operating voltages of 4.0 V and 5.3 V at the luminance of 1000 and 10 000 cd m⁻², respectively. For a green phosphorescent device, approximately 1.2-fold improvement in external quantum efficiency was obtained compared to the conventional structure. We attributed the improved performance to dipole–dipole interactions at the sulfide-organic interface.

---

Introduction

Organic light-emitting diodes (OLEDs) have attracted considerable interest over the last decade due to their potential use in general lighting and active-matrix displays.^1,2 So far, small OLED panels are already being utilized for cellular phones. However, many technical challenges exist for mass production of large OLED panels.^3,4 An important limitation hindering the commercialization of large area displays is related to their intrinsic driving method. The development of an inverted OLED (IOLED) is one of the efforts to match OLEDs to the existing n-type amorphous silicon thin film transistor (TFT) technology.^5–7 When the IOLED is combined with an n-type TFT, there is no voltage drop in the OLED device because the bottom cathode of the IOLEDs is directly connected to the n-type TFT drain line and the anode is connected to the source of the TFT.⁸ Moreover, OLEDs with inverted architectures provide longer device lifetime because water- and oxygen-sensitive electron injection materials can be kept beneath the organic and metal layers.^9,10 Nevertheless, IOLEDs still face some problems such as low efficiency level and high operating voltage.^11–13

One of the key factors determining the device characteristics is the charge injection from the conducting electrodes to the organic semiconductors. Efficient charge injection is generally achieved by matching work functions of the electrode of the anode and cathode to the highest occupied molecular orbital (HOMO) levels and lowest unoccupied molecular orbital (LUMO) levels of the organic semiconductors, respectively.^1,14 High-work function metal oxides such as tungsten oxide (WO₃),^15,16 vanadium pentoxide (V₂O₅)^17,18 or molybdenum oxide (MoO₃)^19,20 have been widely used for hole-injection materials. The main challenge in enhancing the performance of IOLEDs is the development of an electron injection layer (EIL) since it requires low-work function and hence reactive electrodes, interface modifiers or buffer layers.^21,22 More recently, some metal oxides have been investigated for efficient electron injection, with some success, and efficient IOLEDs have been realized by employing zinc oxide or tin dioxide.^23,24 However, the metal sulfide, which may play an important role in enhancing electron injection dynamics, has not been examined.

For practical application, it is also important to be able to produce high-efficiency blue emission from an IOLED at high luminance such as 1000 or 5000 cd m⁻².²⁵ Unfortunately, typical IOLEDs encounter significant roll-off (i.e., current efficiency (η_Ce) or external quantum efficiency (η_EQE) drastically drops as the applied voltage or brightness increases) problems, leading to undesired low device efficiency at high luminance, which is unfavourable to their commercial realization for lighting.^26,27 Hence, IOLED devices with mild or even little roll-off are also highly expected.

In this paper, a novel kind of double electron injection layers (D-EILs) composed of metal sulfide and cesium carbonate (Cs₂CO₃)-doped 4,7-diphenyl-1,10-phenanthroline (Bphen) is reported. We demonstrate the effects of the thickness and species of sulfide film on the electroluminescence (EL) characteristics of blue florescent devices. By using (2 nm-thick metal sulfide)/Bphen: Cs₂CO₃ as D-EILs, the low operating voltage, high electron/photon conversion efficiency and high power efficient of the blue electroluminescence device are obtained. Notably, our device exhibits a surprisingly roll-up character and a beyond theoretical limited η_EQE of 6.9% at 1000 cd m⁻², and 7.5% at 10 000 cd m⁻². These peculiarities are discussed from the standpoint of the interface dipole of the metal sulfide. Furthermore, the optimized D-EILs are used as the basis for developing an advanced green phosphorescent IOLED, achieving a current efficiency enhancement of 21.5% and a power efficiency (η_Pe) enhancement of 20.6%, respectively.

Results and discussion

Device architecture

Fig. 1(a) presents a schematic diagram of the device structure constructed with D-EILs by adopting a novel type of metal sulfide [zinc sulfide (ZnS) or copper sulfide (CuS)] layer and a Bphen layer doped with Cs₂CO₃. We compare six metal sulfide-based IOLEDs, with their structural details provided in Experimental section. Fig. 1(b) presents the reference devices (ref. 1 and 2) and Ir(ppy)₃-based IOLED structures. Fig. 1(c) shows the chemical structures of the host 1,4-bis[N-(1-naphthyl)-N′-phenylamino]-4,4′-diamine/9,10-di(2-naphthyl)anthracene (ADN), 4,40-bis(N-carbazolyl)biphenyl (CBP), blue dopant p-bis(p-N,N-diphenylaminostyryl)benzene (DSA-Ph) and green dopant fac-tris(2-phenylpyridine)iridium [Ir(ppy)₃].

Fig. 1 Structures of the conventional OLEDs and IOLEDs. (a) Device scheme of D-EILs-based blue IOLED. (b) Structures of reference devices and Ir(ppy)₃-based IOLED. (c) Chemical structures of the emitting materials used in the EMLs of the OLEDs.

Metal sulfide, Bphen:10 wt% Cs₂CO₃, lithium fluoride (LiF) function as EIL. Bphen, 1,3,5-tris(2N-phenylbenzimidazolyl)benzene (TPBi) function as the electron transport layers (ETL). 1,4-Bis[N-(1-naphthyl)-N′-phenylamino]-4,4′diamine (NPB), 4,4′,4′′-tris-(N-carbazolyl)-triphenylamine (TCTA) function as the hole transport layer (HTL); and the molybdenum oxide (MoO₃) functions as the hole injection layer (HIL).

Injection property of the pristine D-EILs structure

In previous studies, the small atom of Li has been replaced by the larger atom of Cs to serve as the n-type dopant. But, pure and low-melting Cs source is rather difficult to deposit by thermal process because it is extremely reactive in air. However, n-type doping of Cs₂O decomposed by cesium carbonate (Cs₂CO₃) can be easily handled in the fabrication processes.²⁸ Further, Wakimoto et al. has reported that Cs₂CO₃ would decomposed into Cs₂O owing to the heat reaction.²⁹ Therefore, Cs₂CO₃ has been widely incorporated for electron injection and transport enhancement in organic devices. We have investigated the effect of electron injection to the Bphen layer in ZnS, neat Bphen and Cs₂CO₃-doped Bphen thin films. The forward-biased current density–voltage (J–V) characteristics of the EL devices with various EILs are shown in Fig. 2(a). It can be seen that the current density of devices with the Bphen: Cs₂CO₃ layer is indeed dramatically increased with respect to devices without the Bphen: Cs₂CO₃ layer. This can be easily understood from the n-type doping, which can increase the mobility of carriers and lead to a generation of very thin space-charge layers at the contacts associated with efficient injection.

Fig. 2 (a) J–V characteristics of inverted bottom emission OLEDs as a function of the electron injection material. (b) η_Pe–L curves of inverted bottom emission OLEDs as a function of the electron injection material. The device structure: ITO/EIL/Bphen/ADN:3 wt% DSA-ph/NPB/MoO₃/Al.

For the case of devices with the ZnS/Bphen D-EILs, the injection of electrons from ITO into the ETL seemed to be more effective. However, we found that the introduction of the ZnS buffer layer decreased power efficiency of the device with Bphen EIL, as proposed in Fig. 2(b). This is understandable from radiative charge recombination, which was correlated with the accumulation of electron at the ZnS/Bphen interface. Indeed, it is easy to induce the crystallization of electron transport layer by the electron accumulation at interface due to Joule heat during device operation, ultimately degrading the efficiency of devices.^30,31 Thus, if the interface electrons transfer fast enough into ETL under applied voltage, then leakage current can be effectively decreased, possibly resulting in a more balanced injection of electrons and holes. This assumption was confirmed by the performance of the ZnS/Bphen: Cs₂CO₃-based device. In Fig. 2, the device with ZnS/Bphen: Cs₂CO₃ D-EILs showed significant improvements in current density and power efficiency. This clearly proves the advantages of ZnS/Bphen: Cs₂CO₃ D-EILs in electron injection dynamics.

Thickness-dependent characteristics of ZnS film

It has been reported that the insertion of some buffer layers often leads to substantial increase in device operating voltage.³² In this case, the thickness of the used buffer layer has to be exactly controlled. Therefore, the effect of thickness of ZnS layer on the electrical and EL properties of blue IOLEDs with ZnS/Bphen: Cs₂CO₃ D-EILs were systematically investigated. For comparison, we also fabricated inverted reference device (ref. 1) using typical Bphen: Cs₂CO₃ as single-EIL. Fig. 3 shows device performance of Devices A, B, C and ref. 1.

Fig. 3 Device performance of IOLEDs. (a) Current density vs. voltage, (b) power efficiency vs. luminance, (c) external quantum efficiency (open symbols)–luminance–current efficiency (solid symbos), and (d) the EL spectra of reference device and Devices A, B, C and ref. 1 under the luminance of 1000 cd m⁻² (the inset of (a) shows photographs of the lighting images of the blue emission IOLEDs at the same voltage and the inset of (d) shows CIE comparison between Devices A and C, respectively.).

The J–V curves of devices were plotted in Fig. 3(a). It can be clearly observed that the thin ZnS film (2 nm) of D-EILs can significantly improve the electron injection property over the entire range of driven voltage. However, when the thickness of ZnS in D-EILs rose to 8.0 nm, the current density of Device C is dramatically decreased, and even lower than single-EIL device. Thus, it seems that the electron injection dynamics is sensitive to the thickness of the ZnS film, which easily results in difficulties for electron injection from ITO. As expected, the utilization of 10 nm ZnS film does significantly deteriorate the device's electrical properties (see Fig. S1, ESI†) and also minimizes luminance efficiency (Fig. S2 in the ESI†). The main performances of Devices A, B, C, and ref. 1 were summarized in Table 1. It can be seen that the turn-on voltage for Device A is 2.8 V, which is consistent with a recent observation of normal OLED with DSA-ph emitter.³³ The inset in Fig. 3(a) is the lighting images of the IOLEDs at a driving voltage of 3.0 V. Apparently, the Device A with 2 nm-thick ZnS displayed the highest color intensity. Accordingly, it was observed that the Device A achieved a brightness of 1000 cd m⁻² at a voltage slightly above 4.0 V; and the maximum luminance reached 19 250 cd m⁻² at a driving voltage of 6.6 V, much higher than other devices (Fig. S3 in the ESI†). This implies that Device A has enhanced electron injection from the ITO/D-EILs interface, leading to a more balanced charge carriers and carrier recombination.

Table 1 Optoelectric characteristics of reference device and devices with sulfide-based D-EILs

Device performance	Reference device	Sulfide-based D-EIL
	Ref. 1	Device A ZnS (2 nm)	Device B ZnS (5 nm)	Device C ZnS (8 nm)	Device E CuS (2 nm)	Device F CuS (5 nm)
a Measured at luminance of 1 cd m^−2.b Measured at luminance of 1000 cd m^−2.c Measured at luminance of 10 000 cd m^−2.
Turn-on voltage^a [V]	3.1	2.8	2.9	4.1	2.8	2.9
Driving voltage^b [V]	5.1	4.1	4.7	9.8	4.0	4.3
Driving voltage^c [V]	6.9	5.7	6.4	14.1	5.3	5.6
ηEQE^b [%]	4.4	6.9	5.8	5.4	6.7	5.3
ηEQE^c [%]	4.9	7.5	6.3	5.6	7.3	5.9
ηCe^b [cd A^−1]	9.2	14.3	11.8	11.1	14.4	10.6
ηCe^c (cd A^−1)	10.0	15.4	12.7	11.4	15.6	11.8
ηPe^b (lm W^−1)	5.6	10.9	7.7	3.5	11.2	7.7
CIE^b (x, y)	(0.16, 0.35)	(0.16, 0.37)	(0.16, 0.35)	(0.17, 0.35)	(0.17, 0.38)	(0.16, 0.34)

---

Due to the outstanding electron injection properties of Device A, we have obtained a very-high power efficiency of 12.7 lm W⁻¹, which was about twice that of reference device as shown in Fig. 3(b). Even at high luminance, the power efficiency maintained still quite high and was comparable to that of the state of the art DSA-ph-based device,^34–36 reaching 10.9 lm W⁻¹ at 1000 cd m⁻², and 9.5 lm W⁻¹ at 5000 cd m⁻². The high power efficiency should be greatly important in the reduction of power assumption for practical applications. However, that is not the end of the story. An unusual roll-up phenomenon was clearly observed in the external quantum efficiency and current efficiency results as the luminance steadily increased for the blue fluorescent device as shown in Fig. 3(c), which is attributed to the pairing of host and guest energy levels. For a low LUMO/HOMO energy level of 2.9/5.6 eV for DSA-ph compared with 2.8/5.7 eV for ADN, as shown in Fig. S4 of the ESI†, there obviously exist LUMO and HOMO energy level gap of 0.1 eV between DSA-ph and ADN. Therefore, the excitions generated predominantly on the guest at low voltage while more excitions produced on the host as voltage increased, leading to an increasing use of the available recombination sites offered by the host and guest.^37,38

Device A reached the highest efficiency of 13.4 and 14.3 cd A⁻¹, and a beyond theoretical limit η_EQE of 6.5 and 6.9%, at the luminance of 100 and 1000 cd m⁻², respectively. Even at a very-high luminance of 10 000 cd m⁻², Device A showed a surprisingly higher efficiency performance with a η_Ce of 15.4 cd A⁻¹, and a η_EQE of 7.5% (shown in Table 1). Note that the EQE of Device A exceeded the 5% theoretical limit, suggesting this device has a near perfect carrier-injection balance. More importantly, the electron injection material ZnS possessing a high refractive index (∼2.36) reduces non-radiative losses from the in-plane wave vector at the interface between the bottom ITO (cathode) and the electron transport layer, increasing the boundary of the theoretical limit of outcoupling efficiency for planar OLEDs.

Fig. 3(d) presents the EL spectra at 1000 cd m⁻² of the four devices. The EL emission of Device A was a blue color with Commission Internationale d'Eclairage (CIE) x, y color coordinates of (0.16, 0.37), corresponding to a dominant emission with a peak emission at λ_max = 468 nm accompanied by a vibronic side-band at 500 nm. And it was found that the EL spectrum of Device A revealed almost identical behavior to Devices B and ref. 1 as shown in Fig. 3(d). However, it is worth noting that the side-band emission from Device C, despite of weak with respect to the major peak in the blue, has a dramatic effect on the CIE coordinates (shown in inset of Fig. 3(d) and Table 1). The thicker ZnS resulted in slightly color change in CIE, from (0.16, 0.37) to (0.17, 0.35), stemming from the difference in optical transmission through the ZnS film. At the visible spectral range, the bare ITO-glass with 2 nm yielded a transmittance of 76% at 468 nm while ITO-glass with 8 nm ZnS exhibited an improved transmittance of 78% at 468 nm (see Fig. 4(a)). In addition, our devices exhibited excellent chromatic stability. For instance, the blue emissions in the Device A were almost constant with increasing current density as shown in Fig. S5 (in the ESI†).

Fig. 4 (a) Transmittance spectra and (b) atomic force microscope images of ZnS thin films with various deposition thickness on ITO glass. Inset of (a) shows magnified transmittance band edge between 400 and 500 nm.

To confirm how the ZnS layer onto ITO acts and to elucidate the great improvement in device performance. We have performed a combined atomic force microscopy (AFM)-X-ray diffraction (XRD) technique to investigate the different surface features of ZnS thin films deposited on ITO glass substrates. The XRD profiles for the ITO-sulfide (ZnS) composites with different thickness of binary sulfide film are given in Fig. S6 (in the ESI†). These patterns have a strong and sharp peak at 2θ = 21.6°, which can be indexed as the (222) reflection of metal sulfides. Besides, the ITO/sulfide composites show enhanced XRD patterns to the blank ITO at 2θ = 37.6°, which are coordinated with that of sulfide reported by Chen et al.^39,40 These confirmed the existence of sulfur ion.

The morphologies of the ITO/ZnS surfaces were further studied using AFM, as shown in Fig. 4(b). The ITO film surface yielded a root mean square (RMS) roughness of 1.52 nm, implying that the surface is not an absolute plane. This may result from the sputter deposition process of commercial ITO. The deposition of ZnS with 2 nm on the ITO surface forms a nearly homogeneous film with RMS roughness decreasing from 1.52 nm to 1.25 nm, referring that metal sulfide molecules are prone to fill in the concave sites of ITO surface. As the ZnS thickness was increased from 5 nm to 8 nm, the grain structures gradually appeared and ITO/ZnS surfaces became rough with many serrate edges and some crystals grew in the whole area, demonstrating that the sulfide molecules are not in the form of a neat layer but partially penetrate into ITO. Therefore, 2 nm-thick ZnS buffer layer greatly improves the stability of the morphology for the following film, which may prevent the degradation process (dark spot formation and growth) of IOLEDs under operating conditions.⁴¹ This assumption was confirmed by the lifetime performance of the Devices A, ref. 1 and 2. Fig. S7 of the ESI† shows the time evolution of the luminance under conditions of dry-nitrogen (<0.1 ppm H₂O and O₂) atmospheres. Devices are operated at room temperature and a constant current density for initial luminance of 3000 cd m⁻². The results indicate that the operation lifetime of Devices A and ref. 1 with inverted architectures are significantly improved. Especially for the ZnS-based Device A, operation lifetime of approximately 468 min declined to 80%, more than 2 times higher than that of the ref. 2 with normal structure.

To further explore the role of metal sulfide in charge injection, we characterized the ZnS film by ultraviolet photoelectron spectroscopy (UPS). Fig. S8 of the ESI† shows the kinetic energy spectra of the secondary electrons escaped from 2 nm-thick metal sulfide. The calculated work function value derived from the cut-off spectra was 3.78 eV for ZnS. The lowered Fermi level of ZnS reduced the energy barrier between the ITO and Bphen: Cs₂CO₃ layer, as illustrated in Fig. 5(a), which resulted in a more efficient electron injection, thus enhancing the IOLED performance. In addition, the interface-state played an important role, such as occupation of states on one side or another of the interface, and leaded to the formation of the dipole. Therefore, we discuss the thickness effect of sulfide film from the origin of dipole–dipole interactions at metal sulfide-organic interfaces, as shown in Fig. 5(b). The dipole would easily diffuse into the doped Bphen layer, then quickly dissociated under the influence of electric field (Fig. 5(b) (left)), resulting in higher current densities. As the film thickness increased, more dipoles produced in metal sulfide-organic interfaces because of the favorable interface dipole layer formation. Parts of dipoles dissociated by electric field increased electron injection, but some of them were quenched through aggregation or other ways due to non-efficient injecting between adjoining dipoles. Even, high dipole density hindered more dipoles dissociation, thus less electrons were injected into Bphen resulting in lower current densities (Fig. 5(b) (right)).

Fig. 5 (a) The energy level alignment in ZnS-based device. (b) Schematic illustration of electron-injecting behaviour at low (left) and high (right) dipole density.

A robust CuS-based D-EILs structure

In recent times, CuS has been extensively used as an electrode material for lithium ion batteries due to their good electronic conductivity and high energy capacity.^42–44 Several attempts have been made to explore CuS as a cathode material for recharge battery and comparable recharge ability was obtained up to 1000 cycles.^45,46 Moreover, CuS has been used as a counter electrode for photo electrochemical cells due to their promising redox reaction with polysulfide electrolyte.^47–49 However, there is no clear statement about CuS as an electron injection material for OLEDs, since its electron-injecting property is not much impressive.

Interestingly, metal sulfide-based D-EILs structure shows enhanced electron injection in electroluminescent devices. As mentioned above, the ZnS/Bphen: Cs₂CO₃ D-EILs with suitable thickness could significantly decrease the leakage current, enhancing the luminance efficiency and possibly resulting in a more balanced injection of electrons and holes. Here, CuS-based IOLEDs with D-EILs were fabricated, and the EL performance was shown in Fig. 6, respectively. The results showed higher current densities and exceptionally high efficiencies compared to reference device, as expected. At a voltage of 5.0 V, the current density of IOLEDs decreased, with values of 46.0, 37.5 and 8.9 mA cm⁻² for Devices E, F and ref. 1, respectively (shown in inset of Fig. 6(a)). The improved current density results in a reduction in the driving voltage of Device E by using a 2 nm-thick CuS. For instance, at a certain luminance of 1000 cd m⁻², the driving voltage of Device E decreased from 5.1 V to 4.0 V (Fig. S9, ESI†). Meanwhile, a lower operation voltage of 5.3 V at 10 000 cd m⁻² was obtained (shown in Table 1). Due to low operating voltage, Device E indicated very-high power efficiency (12.3 lm W⁻¹ at 100 cd m⁻², and 11.2 lm W⁻¹ at 1000 cd m⁻²), as illustrated in Fig. 6(a). Importantly, the above mentioned efficiency roll-up characteristic was also observed for the CuS-based device, see Fig. 6(b) for the external quantum efficiency-luminance-current efficiency curves. It is noteworthy that Device E showed a higher η_EQE of 6.7% and a higher η_Ce of 14.4 cd A⁻¹ at 1000 cd m⁻², afterwards increased to 7.3% and 15.6 cd A⁻¹ at 10 000 cd m⁻², respectively. In fact, in the case of sulfide-based device, the external quantum efficiency steadily increased with luminance suggesting that the rate of hole–electron pair recombination continuously improves with the increasing electric field intensity. This improvement in radiative charge recombination therefore helps to suppress the efficiency roll-off. Simultaneously, Device E gave blue electroluminescence centred at 468 nm. Devices F and ref. 1 exhibited the same EL spectra with an identical peak wavelength as shown in Fig. S10 (in the ESI†), which is independent of the increasing driving voltage. It is worth to note that the side-band emission from CuS-based device, despite of weak with respect to the major peak in the blue, has heavily effect on the color change of the device, with results in inset of Fig. S10a (ESI†). The color change are attributed to the thick CuS film due to relatively strong absorption in visible light (see Fig. S10b, ESI†).

Fig. 6 (a) Power efficiency–luminance (and inset: current density–voltage) for Devices E, F and ref. 1; (b) current efficiency and external quantum efficiency plotted against luminance curves for Devices E, F and ref. 1.

High-efficiency phosphorescent IOLED based on D-EILs structure

Although blue fluorescent OLED has been investigated to demonstrate the D-EILs based IOLED, other colors can also be realized by varying the emitter. For instance, a green device has been produced with a similar structure, but with an Ir-complex emissive layer instead. Previous studies have demonstrated that small amount of phosphorescent molecular doped in CBP can facilitate electron transfer.^50,51 Thus, we use ZnS/Bphen: Cs₂CO₃ with 5 nm-thick ZnS as D-EILs to achieve better carrier balance for radiative recombination. Additionally, the introduction of TCTA and TPBi layers did not only transport carriers effectively, but play a significant role in blocking carriers, by which confine exciton in EML for avoiding the energy outflow.⁵² Similarly, the reference device ref. 3 with a typical normal structure was also demonstrated.

As shown in Fig. S11 (in the ESI†), the maximum current efficiency and the power efficiency of the fabricated devices were 76.4 (Device G), 62.9 cd A⁻¹ (ref. 3), and 65.7 (Device G), 54.5 lm W⁻¹ (ref. 3), which were enhanced by 21.5% and 20.6% respectively. Remarkably, the efficiency-luminance characteristics of Device G remained stable until very high luminance were reached (76.4 cd A⁻¹ at 300 cd m⁻² vs. 63.9 cd A⁻¹ at 3000 cd m⁻²). This indicates that the efficiency possesses very low efficiency roll-off under high luminance. Furthermore, we observed that the peak external quantum efficiency (ca. 22.5%) of Device G showed ca. 1.2 times higher than that of ref. 3 (ca. 18.3%) (Fig. S12, ESI†), which can be attributed to strong electron-injecting ability of D-EILs. The EL spectra of devices were shown in Fig. S13 (in the ESI†), and they exhibited almost identical spectra with a peak at 516 nm.

Conclusions

In summary, the D-EILs consisting of a novel type of metal sulfide layer and a Bphen layer doped with Cs₂CO₃ was used to enhance electron injection and mitigate the operation voltage in IOLEDs. We have found that the D-EILs structure provided superior electron injection performance and carrier balance in IOLEDs. Importantly, the IOLED performance was clearly dependent on the thickness and species of metal sulfide (e.g., ZnS, CuS). Blue florescent device based on (2 nm-ZnS)/Bphen: Cs₂CO₃ D-EILs exhibited a surprising η_EQE of 7.5%, a satisfactory power efficiency of 12.7 lm W⁻¹ and a very low turn-on voltage (at 1 cd m⁻²) of 2.8 V. Meanwhile, (2 nm-CuS)/Bphen: Cs₂CO₃-based device also presented extremely low driving voltage (4.0 V at 1000 cd m⁻² and 5.3 V at 10 000 cd m⁻²) and higher efficiency (7.3% and 15.6 cd A⁻¹ at 10 000 cd m⁻², respectively). The formation of a favorable interfacial dipole layer at the metal sulfide-organic interface was shown to be the main reason for the improved performance of IOLED with D-EILs. In addition, our device presented an unusual roll-up peculiarity (a η_EQE increasing from 6.9 to 7.5% and a η_Ce increasing from 14.3 to 15.4 cd A⁻¹ with the rise in luminance from 1000 to 10 000 cd m⁻²). More importantly, Ir(ppy)₃-based green IOLED achieved an approximately 1.2-fold luminescence efficiency improvement in comparison to the conventional normal OLED due to the advanced injection strategies of D-EILs. This demonstration of effective EL offers a scope for developing this unique class of D-EILs into efficient AMOLED display.

Experimental section

Materials

All chemicals and regents in this work were used as received from commercial sources without purification. The small-molecular materials were obtained from e-Ray Optoelectronics Corp. (Taipei, Taiwan). Molybdenum(VI) oxide (99.99%), zinc sulfide (powder, 10 μm, 99.99% trace metals basis) and copper sulfide (powder, 100 mesh, ≥99% trace metals basis) were purchased from Sigma-Aldrich. Indium tin oxide (ITO, 15 Ω per sheet, 150 nm)-coated glass substrates were ordered from CSG Holding Co. Ltd (China).

Device fabrication

The OLEDs were fabricated by conventional vacuum deposition of the organic layers, metal sulfide and Al electrode onto an ITO-coated glass substrate under a base pressure lower than 5 × 10⁻⁴ Pa. Devices A, B, C and D have the following structure: indium tin oxide (ITO) cathode/ZnS (χ nm)/Bphen:10 wt% Cs₂CO₃ (30 nm)/Bphen (10 nm)/ADN:3 wt% DSA-ph (25 nm)/NPB (40 nm)/MoO₃ (5 nm)/Al anode (100 nm), where χ are 2, 5, 8 and 10, respectively. The structure of Devices E and F is ITO cathode/CuS (2 or 5 nm)/Bphen:10 wt% Cs₂CO₃ (30 nm)/Bphen (10 nm)/ADN:3 wt% DSA-ph (25 nm)/NPB (40 nm)/MoO₃ (5 nm)/Al anode (100 nm). The structure of reference Device 1 (ref. 1) is ITO cathode/Bphen:10 wt% Cs₂CO₃ (30 nm)/Bphen (10 nm)/ADN:3 wt% DSA-ph (25 nm)/NPB (40 nm)/MoO₃ (5 nm)/Al anode (100 nm). The structure of reference Device 2 (ref. 2) is ITO anode/MoO₃ (5 nm)/NPB (40 nm)/ADN:3 wt% DSA-ph (25 nm)/Bphen (10 nm)/Bphen:10 wt% Cs₂CO₃ (30 nm)/LiF (0.3 nm)/Al cathode (100 nm). The structure of reference Device 3 (ref. 3) is ITO anode/MoO₃ (5 nm)/NPB (25 nm)/TCTA (10 nm)/CBP:10 wt% Ir(ppy)₃ (20 nm)/TPBi (30 nm)/LiF (0.3 nm)/Al cathode (100 nm). While Device G have the following structure: ITO cathode/ZnS (5 nm)/Bphen:10 wt% Cs₂CO₃ (20 nm)/TPBi (10 nm)/CBP:10 wt% Ir(ppy)₃ (20 nm)/TCTA (10 nm)/NPB (25 nm)/MoO₃ (5 nm)/Al anode (100 nm). All devices were grown pre-patterned ITO with an active area of 4 mm² defined by the overlap of electrodes. Prepared ITO-coated glass substrates were cleaned by detergent, de-ionized water, acetone, and isopropanol. Immediately prior to loading into a custom-made high vacuum thermal evaporation chamber, the substrates were treated to a UV-ozone environment for 15 min. Then, the entire organic layers, metal sulfide and Al electrode were deposited successively without exposure to the atmosphere. The deposition rates for the organic materials, metal sulfide, LiF, and Al were typically 0.1, 0.02, 0.01 and 0.5 nm s⁻¹, respectively. The layer thickness was controlled in situ using a quartz crystal monitor.

Device measurements

The current–voltage–luminescence characteristics were measured by a Keithley 2400 source meter and a PR-650 Spectra Colorimeter. The luminance and spectra of each device were measured in the direction perpendicular to the substrate. The surface morphology of the ZnS layer coated on the ITO glass was investigated by the AFM technique in contact mode using the Seiko instrument SPA 400 AFM system. And the optical property was measured with an ultraviolet-visible-near infrared spectrophotometer (U-3900H, Hitachi). Moreover, the crystallization process of metal sulfide thin film was checked by XRD analysis with a Rigaku D/MAX 2550V X-ray diffractometer using Cu Kα radiation. Ultraviolet photoelectron spectroscopy (UPS) spectra were collected and analyzed in an Escalab 250Xi Surface Analysis System (Thermo Scientific). A UV lamp emitted He I (21.2 eV) radiation lines were used in UPS measurements.

Acknowledgements

This work was financially supported by the “973” program (2015CB655005), the National Natural Scientific Foundation of China (61136003, 51505270, 61565003), and the Science and Technology Committee of Shanghai (15590500500).

Notes and references

1. H. Fukazawa, K. Morii, M. Hasegawa, Y. Arimoto, T. Kamada, T. Shimizu and T. Yamamoto, Appl. Phys. Express, 2014, 7, 082104.
2. E. Najafabadi, K. A. Knauer, W. Haske and B. Kippelen, Org. Electron., 2013, 14, 1271.
3. H. Lee, H. Ahn, C. Lee and J. Kwak, J. Korean Phys. Soc., 2012, 61, 1536.
4. L. Ding, S. C. Dong, Z. Q. Jiang, H. Chen and L. S. Liao, Adv. Funct. Mater., 2014, 25, 645.
5. Y. Lee, J. Kim, S. Kwon, C.-K. Min, Y. Yi, J. W. Kim, B. Koo and M. Hong, Org. Electron., 2008, 9, 407.
6. C. H. Park, H. J. Lee, J. H. Hwang, K. N. Kim, Y. S. Shim, S. G. Jung, C. H. Park, Y. W. Park and B. K. Ju, ACS Appl. Mater. Interfaces, 2015, 7, 6047.
7. C. H. Chang, M. K. Hsu, S. W. Wu, M. H. Chen, H. H. Lin, C. S. Li, T. W. Pi, H. H. Chang and N. P. Chen, Phys. Chem. Chem. Phys., 2015, 17, 13123.
8. H.-H. Hsieh, T.-T. Tsai, C.-Y. Chang, H.-H. Wang, J.-Y. Huang, S.-F. Hsu, Y.-C. Wu, T.-C. Tsai, C.-S. Chuang, L.-H. Chang and Y.-H. Lin, Dig. Tech. Pap.–Soc. Inf. Disp. Int. Symp., 2010, 41, 140.
9. P. S. Wang, I. W. Wu and C. I. Wu, J. Appl. Phys., 2010, 108, 103714.
10. C. W. Chen, C. L. Lin and C. C. Wu, Appl. Phys. Lett., 2004, 85, 2469.
11. S. Höfle, A. Schienle, M. Bruns, U. Lemmer and A. Colsmann, Adv. Mater., 2014, 26, 2750.
12. K. Zhang, C. Zhong, S. Liu, A. H. Liang, S. Dong and F. Huang, J. Mater. Chem. C, 2014, 2, 3270.
13. M. Thomschke, S. Hofmann, S. Olthof, M. Anderson, H. Kleemann, M. Schober and K. Leo, Appl. Phys. Lett., 2011, 98, 083304.
14. H. Sasabe, N. Toyota, H. Nakanishi, T. Ishizaka, Y. J. Pu and J. Kido, Adv. Mater., 2012, 24, 3212.
15. J. Meyer, S. Hamwi, T. Bülow, H. H. Johannes, T. Riedl and W. Kowalsky, Appl. Phys. Lett., 2007, 91, 113506.
16. J. Meyer, T. Winkler, S. Hamwi, S. Schmale, H. H. Johannes, T. Weimann and T. Riedl, Adv. Mater., 2008, 20, 3839.
17. J. Wu, J. Hou, Y. Cheng, Z. Xie and L. Wang, Semicond. Sci. Technol., 2007, 22, 824.
18. H. Lee, Y. K. Kim and C. Lee, Dig. Tech. Pap.–Soc. Inf. Disp. Int. Symp., 2012, 43, 1477.
19. S. Tokito, K. Noda and Y. Taga, J. Phys. D: Appl. Phys., 1996, 29, 2750.
20. T. Schwab, S. Schubert, M. Thomschke, M. Fröbel, L. Müller-Meskamp, K. Leo and M. C. Gather, Dig. Tech. Pap.–Soc. Inf. Disp. Int. Symp., 2013, 44, 600.
21. J. Liu, X. K. Wu, X. D. Shi, J. Wang, Z. Y. Min, Y. Wang, M. J. Yang and G. F. He, ACS Appl. Mater. Interfaces, 2015, 7, 6438.
22. S. H. Ouyang, Y. T. Xie, D. P. Wang, D. L. Zhu, X. X. Xu, T. Tan and H. H. Fong, J. Disp. Technol., 2015, 11, 273.
23. S. Höfle, A. Schienle, M. Bruns, U. Lemmer and A. Colsmann, Adv. Mater., 2014, 26, 2750.
24. H. Lee, C. M. Kang, M. Park, J. Kwak and C. Lee, ACS Appl. Mate. Interfaces, 2013, 5, 1977.
25. J. H. Jou, P. C. Chen, M. C. Tang, Y. S. Wang, C. H. Lin, S. H. Chen, C. C. Chen, C. C. Wang and C. T. Chen, J. Photonics Energy, 2012, 2, 021208.
26. N. C. Giebink and S. R. Forrest, Phys. Rev. B: Condens. Matter Mater. Phys., 2008, 77, 235215.
27. J. Chen, C. Shi, Q. Fu, F. Zhao, Y. Hu, Y. Feng and D. Ma, J. Mater. Chem., 2012, 22, 5164.
28. S. Y. Chen, T. Y. Chu, J. F. Chen, C. Y. Su and C. H. Chen, Appl. Phys. Lett., 2006, 89, 053518.
29. T. Wakimoto, Y. Fukuda, K. Nagayama, A. Yokoi, H. Nakada and M. Tsuchida, IEEE Trans. Electron Devices, 1997, 44, 1245.
30. H. You, Y. Dai, Z. Zhang and D. Ma, J. Appl. Phys., 2007, 101, 026105.
31. S. Urakawa, S. Tomai, Y. Ueoka, H. Yamazaki, M. Kasami, K. Yano and Y. Uraoka, Appl. Phys. Lett., 2013, 102, 053506.
32. C. C. Yap, M. Yahaya and M. M. Salleh, Curr. Appl. Phys., 2008, 8, 637.
33. B. Liu, M. Xu, L. Wang, H. Tao, Y. Su, D. Gao, J. Zou, L. Lan and J. Peng, ECS J. Solid State Sci. Technol., 2013, 2, R258.
34. X. Wu, W. Bi, Y. Hua, J. E. Sun, Z. Xiao, L. Wang and S. Yin, Appl. Phys. Lett., 2013, 102, 243302.
35. K. Guo, J. Zhang, T. Xu, X. Gao and B. Wei, J. Disp. Technol., 2014, 10, 642.
36. C. Li, T. Tsuboi and W. Huang, Opt. Mater., 2014, 36, 1720.
37. J.-H. Jou, P.-C. Chen, M.-C. Tang, Y.-S. Wang, C.-H. Lin, S.-H. Chen, C.-C. Chen, C.-C. Wang and C.-T. Chen, J. Photonics Energy, 2012, 2, 021208.
38. K. Guo, J. Zhang, T. Xu, X. Gao and B. Wei, J. Disp. Technol., 2014, 10, 642.
39. L. Chen, J. D. Liu and S. Q. Zhang, J. Inorg. Mater., 2013, 28, 1127.
40. Z. Li, L. Mi, W. Chen, H. Hou, C. Liu, H. Wang and C. Shen, CrystEngComm, 2012, 14, 3965.
41. R. Phatak, T. Y. Tsui and H. Aziz, J. Appl. Phys., 2012, 111, 054512.
42. A. Débart, L. Dupont, R. Patrice and J. M. Tarascon, Solid State Sci., 2006, 8, 640.
43. T. Ohzuku and R. J. Brodd, J. Power Sources, 2007, 174, 449.
44. J. S. Chung and H. J. Sohn, J. Power Sources, 2002, 108, 226.
45. Y. Wang, X. Zhang, P. Chen, H. Liao and S. Cheng, Electrochim. Acta, 2012, 80, 264.
46. I. Exnar and J. Hep, J. Power Sources, 1993, 44, 701.
47. S. S. Kalanur, S. Y. Chae and O. S. Joo, Electrochim. Acta, 2013, 103, 91.
48. C. J. Raj, K. Prabakar, A. D. Savariraj and H. J. Kim, Electrochim. Acta, 2013, 103, 231.
49. C. J. Raj, B. C. Kim, W. J. Cho, W. G. Lee, Y. Seo and K. H. Yu, J. Alloys Compd., 2014, 586, 191.
50. Z. Gao, F. Wang, K. Guo, H. Wang, B. Wei and B. Xu, Opt. Laser Technol., 2014, 56, 20.
51. Y. Lv, P. Zhou, N. Wei, K. Peng, J. Yu, B. Wei and C. Li, Org. Electron., 2013, 14, 124.
52. H. Sasabe, J. I. Takamatsu, T. Motoyama, S. Watanabe, G. Wagenblast, N. Langer, O. Molt, E. Fuchs, C. Lennartz and J. Kido, Adv. Mater., 2010, 22, 5003.

---

Footnote

† Electronic supplementary information (ESI) available: Additional J–V–L, Ce–L–Pe, EQE–J, EL spectra, and XRD patterns. See DOI: 10.1039/c6ra08191f

---