1. Introduction

White organic light emitting devices (WOLEDs) have been subjects of intensive studies and continually improved due to their great potential for power efficient full color display, backlighting, and illumination applications [–20]. Phosphorescent materials have been widely used in WOLEDs [6, 7, 10, 11–13, 19–21], due to their higher efficiency by using both singlet and triplet excitons comparing to fluorescent materials. Recently, Universal Display Corporation has successfully demonstrated a WOLED with a power efficiency of 102 lm/W at 1000 cd/m² by using phosphorescent materials [21].

Generally, an additive mixture of three primary colors (red, green, and blue, RGB) or two complementary colors is required to generate white light. To date, there are many approaches reported to realize WOLEDs, mainly by employing a single emissive layer doped simultaneously with different color light emitting materials [1, 7, 13], or stacked multiple emissive layers in which each layer emits different color light, respectively [2, 3, 6, 9, 11, 17, 19, 20], as well as tandem structures [8, 10]. In the WOLEDs with single emissive layer, relative doping concentrations of different light emitting materials have to be carefully adjusted and precisely controlled to keep an exciton distribution balance among each light emitting material [1, 7, 13]. On the other hand, in the WOLEDs with multiple emissive layers, carrier or/and exciton blocking layers (spacer) are usually inserted between emissive layers to obtain carrier recombination and exciton distribution balances in each emissive layer [6, 11, 19, 20]. Moreover, these WOLEDs show color shift with drive current or/and operating times, which probably results from change of the carrier recombination region, exciton distribution, or/and differential aging of different emitting materials.

Other easy and convenient fabrication techniques to produce WOLEDs with improved color stability are combination of blue devices with color conversion layers (CCLs) [4, 5, 12, 14–16]. Duggal et al. demonstrated illumination quality WOLEDs, with an efficiency of 15 lm/W at 1000 cd/m², by combining polyfluorene-based blue polymer LEDs with CCLs [4, 5]. Based on the same approach, Krummacher et al. obtained solution processed WOLEDs with an improved efficiency of 25 lm/W, by using blue electrophosphorescent emitting materials [12]. Moreover, the light extraction efficiency of the devices was found to be much improved in these reports due to volumetric light scattering of phosphor particles in the CCLs [4, 5, 12]. Mikami reported an interesting WOLED structure, in which emission from the blue devices is laterally transferred to adjacent orange CCLs [14]. Li et al. proposed a WOLED structure with a CCL as hole injection layer, which increases the driving voltage of the devices [15, 16].

To the best of our knowledge, no WOLEDs have been demonstrated combining vacuum deposited blue electrophosphorescent devices with CCLs. In this letter, we report the white organic light-emitting diodes (WOLEDs) combining vacuum deposited blue electrophosphorescent devices with red surface CCLs. With an iridium (III) [bis(4,6-di-fluoropheny)-pyridinato-N,C²’] picolinate (FIrpic) doped 4,4’-bis(9-carbazolyl)-2,2’-dimethyl-biphenyl (CDBP) blue electrophosphorescent light emitting layer, and an appropriate red surface CCL containing 4-(dicyanomethylene)-2-tert-butyl-6-(1,1,7,7-tetramethyljulolidyl-9-enyl)-4H-pyran (DCJTB), the WOLED generate high efficiency and very pure white light with forward viewing peak luminous (power) efficiencies of 18.1 cd/A (9.5 lm/W) and CIE coordinates of (0.32, 0.31), very close to the equal-energy white, respectively. Moreover, the output spectra and CIE coordinates of the WOLED show no significant change at a wide range of current density.

2. Experimental results and discussion

The WOLED consist of a vacuum deposited FIrpic based blue electrophosphorescent device on indium tin oxide (ITO) coated soda lime glass substrate, and a surface CCL containing DCJTB applied to the outside surface of the ITO glass substrate, as shown in Fig. 1(a) and 1(b). Firstly, the DCJTB was thoroughly dispersed in the epoxy resin (Swancor EP-400A/B with a refractive index of 1.53, which is nearly identical to the refractive index of the soda lime glass) with mass concentrations of 0.2%, 0.3%, and 0.4%, respectively. These mixtures were then spin-coated on the outside of the ITO glass substrates respectively. Finally, the CCLs containing DCJTB were cured at 70 °C for 2 hours. The thickness of the cured layers were around 10 μ m as measured by profilemetry.

Following the routine cleaning steps, the ITO glass substrates were transferred into the a multi-source vacuum chamber at a base pressure of around 10^-6 Torr for fabrication of devices with structure of ITO / MoO₃ (3 nm) / NPB (40 nm) / TCTA (10 nm) / CDBP: FIrpic (8%, 30 nm) / TAZ (10 nm) / BPhen (20 nm) / CsCO₃ (1 nm) /Al (100 nm), where MoO₃ as hole injection layer, N,N’-bis-(1-naphthl)-diphenyl-1,1’-biphenyl-4,4’-diamine (NPB) as the hole transport layer, FIrpic doped CDBP as blue light emitting layer [22], 4,4’,4”-tri(N-carbazolyl)-triphenylamine (TCTA) and 3-(4-biphenylyl)-4-phenyl-5-(4-tert-butylphenyl)-1,2,4-triazole (TAZ) as the exciton blocking layer, 4,7-diphenyl-1,10-phenanthroline (BPhen) as the electron transport layer, CsCO₃ as the electron injection layer. All material layers in the devices were thermally evaporated in sequence without breaking vacuum. The FIrpic doped CDBP blue light emitting layer was deposited by co-evaporation of FIrpic and CDBP, the evaporation rate of FIrpic and CDBP can be controlled independently by measuring them with two separate quartz thickness monitors, allowing the mass doping ratio to be controlled at 8%.

The devices with active area of 0.36 cm² were characterized immediately after the device fabrication at room temperature in air. The current density - forward viewing luminance -voltage characteristics of the devices were measured using a Keithley SMU 236 source-measure unit and a commercial luminance meter equipped with a calibrated silicon photodiode (ST-86LA, from Photoelectric Instrument Factory of Beijing Normal University, China). The electroluminescent (EL) spectra of the devices were recorded by a fiber spectrometer (Ocean Optics USB 2000). By assuming that the emission pattern of the devices were Lambertian, we could calculate the luminous (power) efficiency from the current density - forward viewing luminance - voltage characteristics of the devices. Moreover, the CIE coordinates were calculated from the EL spectra of the devices. The photoluminescence (PL) and absorption spectra of the CCLs containing DCJTB were recorded by a fluorescence spectrophotometer (Hitachi F-4500) and a UV/VIS/NIR spectrometer (PerkinElmer Lambda 900), respectively.

 

Fig. 1. (a) Blue device structure, (b) WOLED structure, (c) absorption and relative PL spectra of the CCLs, (d) normalized EL spectra, and (e) CIE coordinates and the corresponding photographs of the devices.

Download Full Size | PDF

The absorption and relative PL spectra (excited at 512 nm) of the CCLs containing DCJTB on ITO coated soda lime glass are shown in Fig. 1(c). The CCLs show strong absorption from 450 nm to 550 nm, which matched well with the emission of the FIrpic. At their absorption peak of 512 nm, the CCLs containing 0.2%, 0.3%, and 0.4% DCJTB exhibit absorbance of around 0.40, 0.75, and 1.07, and absorb around 61%, 82%, and 92% optical excitation, respectively. The CCLs emit the red light with a peak at around 580 nm originated from the DCJTB and exhibit a little bathochromic shift with increased DCJTB concentration. The fluorescent quantum efficiencies were estimated to be around 75%, 56%, and 50% for the CCLs containing 0.2%, 0.3%, and 0.4% DCJTB from their relative PL intensity (integrated area of the PL band after correcting the absorption difference at 512 nm) by comparing with that (32%) of a 100 nm thick tris (8-quinolinolato) aluminum (Alq₃) solid thin film evaporated on ITO coated soda lime glass [23]. The color conversion coefficients were evaluated to be approximately 45% for all CCLs in this study from their absorption at 512 nm and fluorescent quantum efficiencies estimated as above mentioned.

The normalized EL spectra of the devices at 10 mA/cm² are shown in Fig. 1(d). The EL spectra of the blue devices shows solely blue emission of FIrpic (with peaks at 470 and 495 nm) corresponding CIE coordinates of around (0.15, 0.30), as shown in Fig. 1(e). The EL spectra of the WOLEDs additionally contain a red component from DCJTB, which absorbs a part of the blue electrophosphorescence of the FIrpic, and convert this to its red fluorescence [7]. The WOLEDs yield a white light emission resulting from final mixture of color-converted red light and unabsorbed blue light. The red emission from the DCJTB grows in relation to the blue emission from the FIrpic with increased DCJTB concentration. By simply varying the concentration of the DCJTB in the CCLs, the color of the WOLEDs can be adjusted significantly.

The CIE coordinates and the corresponding photographs of the operating devices are shown in Fig. 1(e). The CIE coordinates of the WOLEDs shift from (0.25, 0.31) of blueish white to (0.40, 0.32) of reddish white with increased DCJTB concentration from 0.2% to 0.4% in the CCLs. A very pure white color of (0.32, 0.31), very close to the equal-energy white (0.33, 0.33), is obtained with a DCJTB concentration of 0.3% in the CCL.

The normalized EL spectra and CIE coordinates of the WOLEDs with a 0.3% DCJTB in the CCL at several current densities are shown in Fig. 2. Obviously, there is no significant EL spectra change or color shift [CIE coordinate migration is less than (±0.004; ±0.004).] with the current density (from 1 to 100 mA/cm²) for the devices, indicating the WOLEDs demonstrated here have excellent color stability.

 

Fig. 2. Normalized EL spectra and CIE coordinates (inset) of the WOLED with a 0.3% DCJTB in the CCL at several current densities.

Download Full Size | PDF

The typical current density - voltage characteristics of the blue electrophosphorescent devices and the forward viewing luminance - voltage characteristics of the devices are shown in Fig. 3(a). The WOLEDs, consisting of the blue electrophosphorescent devices with the CCLs, exhibit nearly similar current density - voltage characteristics and a little lower forward viewing luminance comparing to the blue electrophosphorescent devices. For example, the blue devices, WOLEDs with CCLs containing 0.2%, 0.3%, and 0.4% DCJTB exhibit forward viewing luminance of around 1300 cd/m², 1040 cd/m², 770 cd/m², and 570 cd/m² at around 9 V and 10 mA/cm², though all devices exhibit a turn on voltage of around 5 V for 1 cd/m².

 

Fig. 3. (a) Typical current density - voltage characteristics of the blue electrophosphorescent devices and forward viewing luminance - voltage characteristics, and (b) efficiencies of the devices.

Download Full Size | PDF

The device efficiencies as a function of current densities are shown in Fig. 3(b). The blue electrophosphorescent devices show a peak luminous (power) efficiency of 25.4 cd/A (13.3 lm/W) at 6 V (41.5 cd/m² and 0.16 mA/cm²), which is among the best reported results [22, 23–27]. The WOLEDs, with CCLs containing 0.2%, 0.3%, and 0.4% DCJTB, show peak luminous (power) efficiencies of 23.0 cd/A (13.1 lm/W), 18.1 cd/A (9.5 lm/W), and 10.3 cd/A (5.4 lm/W), respectively. The performance of the devices was summarized in Table 1.

Except the lower fluorescent quantum efficiencies of the CCLs containing DCJTB, the main reason for lower luminance and efficiency of the WOLEDs comparing to the blue electrophosphorescent devices are due to the planar structure of the surface CCLs. It should be noted that although the CCLs can absorb both the forward-viewing and glass substrate waveguided blue emission, however due to the large difference of refractive index between the CCLs (1.53) and air, the absorbed light is reemitted isotropically and around 80% of the reemitted light is totally reflected and waveguided in both the glass substrate and the CCLs.

Table 1. Performance of the devices.

View Table

3. Summary

In summary, we demonstrated easy fabricated WOLEDs with improved color stability by combining vacuum deposited blue electrophosphorescent devices with red surface CCLs. The WOLEDs with an appropriate CCL exhibit a peak luminous (power) efficiency of 18.1 cd/A (9.5 lm/W) and very pure white light emission with CIE coordinates of (0.32, 0.31). However, the efficiency of these WOLEDs is substantially lower than those of state-of-the-art WOLEDs reported [5, 6, 10–12, 19–21]. There should be still plenty of room in improving efficiencies of the vacuum deposited FIrpic-based blue electrophosphorescent devices, e.g. by using effective host materials with larger triplet energy or electrically doped charge transport layers [24–27, 28]. Moreover, to improve the light extraction efficiency of the WOLEDs demonstrated here, further efforts, e.g. employing light extraction enhancing microlenses or volumetric light scattering particles [4–6, 12, 19, 29], are required.