1. Introduction

Top-emitting organic light-emitting devices (TEOLEDs) are of increasing interest since their architectures allow light outcoupling from the top electrode, which can provide a high aperture ratio and a high resolution in active matrix OLEDs and can also realize microdisplays on Si substrates. In the fabrication of TEOLEDs, metal films are usually used as the semitransparent cathode because of the advantage of no damage to the organic layers during the thermal deposition [1–3]. However, due to the high reflection of the cathode and the anode with metals, microcavity effects exit inevitably in TEOLEDs, resulting in a strong viewing angle dependence. The high reflection of the metal cathode also leads to a strong reflection of ambient light with a result of a low pixel contrast ratio (PCR) in TEOLEDs. So the metal cathode with a low reflection is needed to acquire both a weak viewing angle dependence and a high PCR. Besides, a proper light outcoupling layer can further reduce the cathode reflectivity and restrain microcavity effects with results of the improvements of the viewing angle and the PCR [4–5].

In this paper, instead of the widely used AlAg cathode, Samarium (Sm) is utilized as the cathode because Sm has a lower reflectivity than Ag (Here, the Al film of the AlAg bilayer with a thickness of ~1 nm can be ignored) in the visible area, as compared in Fig. 1(a). In addition, Sm has a low work function of only ~2.7 eV, which may be more suitable to be the semitransparent cathode. However, to obtain a TEOLED with a long lifetime and stable EL performances, the Sm/Ag bilayer is adopted with another Ag layer covered onto the Sm layer. So we fabricate green TEOLEDs with the Sm/Ag bilayer and the 2,9-dimethyl-4,7-diphenyl-1,10-phenanthroline (BCP) film as the semitransparent cathode and the light outcoupling layer, respectively. Here, BCP is used due to a low vacuum evaporation temperature less than 100 degree and the absorption transparence in the visible area. As a result, the electroluminescent (EL) spectra shows only 8 nm blueshift by changing the viewing angle from 0 to 80° along with the introduction of the BCP layer. Compared with the common AlAg cathode, the PCR based on the SmAg cathode and the BCP outcoupling layer increases by about 100%.

2. Experimental details

SiO₂-covered silicon substrates are cleaned with acetone and ethanol by using an ultrasonic bath, rinsed with deionized water, and then dried in an oven. The TEOLED architectures are Ag (100 nm)/Ag₂O (UV-ozone for 30 s)/4,4′, 4″-tris(3-methylphenylphenylamino) triphenylamine (m-MTDATA, 45 nm)/4,4′-bis[N-(1-naphthyl-1-)-N-phenyl-amino]-biphenyl (NPB, 5 nm)/tris-(8-hydroxyl-quinoline) aluminum (Alq₃, 50 nm)/LiF (1 nm)/cathode [the Al (1 nm)/Ag (22 nm) cathode for device A, the Sm (11 nm)/Ag (12 nm) cathode for device B, and an 97 nm-thick BCP layer onto the SmAg cathode for device C], as shown in Fig. 1(b). Here, m-MTDATA, NPB, and Alq₃ are utilized as a hole-injection layer, a hole-transporting layer, and an electron-transporting and emissive layer, respectively. All depositions are in a high vacuum above 10^-6 Torr with a rate of 0.1-0.3 nm/s. The current-voltage-luminance characteristics are measured with a programmable Keithley Model 2400 and a Photo-Research PR-650 SpectraScan Colorimeter in room-temperature air. Optical constants of the organic materials used in the work are measured with variable angle spectroscopic ellipsometry. The transmittivity and the reflectivity curves of the devices are calculated with the transfer matrix theory [6].

3. Results and discussion

From the current density-voltage (J-V) characteristics in Fig. 2(a), we find that the electron injection from the SmAg cathode (devices B and C) is much worse than from the AlAg one (device A), mainly attributed to a larger energy barrier between LiF/Sm than that between LiF/Al. Moreover, the luminance-voltage (L-V) characteristic of device A is higher than that of device B, which is attributed to both a fine electron injection and a strong microcavity effect in device A. From Fig. 2(b), the L in device A is still higher than that in device B under the same injected J because of a high reflectivity R(λ) of the AlAg cathode over the SmAg one [See Fig. 3(a)], corresponding to a strong microcavity effect. Here, the R(λ) values are calculated with the transfer matrix theory with the effect of organic layers under consideration. This phenomenon can be further explained with the resonant emission enhancement factor G_cav(λ) [7], which is used to describe the magnified light from the cavity at the wavelength λ:

 

Fig. 1. (a) Comparison of reflective curves for the 23 nm-thick Sm and Ag film, and (b) schematic illustration of devices A, B, and C.

Download Full Size | PDF

Here, ξ/2 is a normalized antinode enhancement factor [8], τ_cav and τ are lifetimes of the molecular excited state in the cavity and in the free space, respectively, R ₁(λ) is the reflectivity of the anode, and R ₂(λ) and T ₂(λ) are the effective reflectivity and transmittivity of the cathode, respectively. The G_cav(λ) values are calculated with combined T(λ) [Fig. 3(a)], R(λ) of both electrodes [Fig. 3(a) and (b)] and the formula (1). The dependence of G_cav(λ) on wavelength for devices A, B, and C is shown in Fig. 3(c). G_cav(λ) of device A is (3.67-37.0)*Z in the range of 380–780 nm, much larger than (4.1–10.7)*Z of device B and (3.3–9.63)*Z of device C, in which (ξ/2)*(τ_cav/τ) is supposed to be a same constant Z in all devices. A high magnified coefficient in device A also leads to a large luminous efficiency (η) of 10.43 cd/A at 6 V, as shown in Fig. 2(c), which is much larger than 5.94 cd/A at 8 V in device B and 7.75 cd/A at 8 V in device C.

 

Fig. 2. (a) (Color online) L-J-V, (b) L-J, and (c) η-V characteristics of devices A, B, and C.

Download Full Size | PDF

Compared with device B, the L and η of device C exhibit an enhancement factor of ~1.4, reaching 114 940 cd/m² at 15 V and 7.75 cd/A at 8 V, respectively. The G_cav(λ) curves of devices B and C in Fig. 3(c) can be referred to give a further explanation about this phenomenon. The G_cav(λ) value of device C is larger than that of device B in the wide range of 440–700 nm that is benefical to the enhancement of the device brightness (L∝V(λ)*f(λ)dλ. Here, f(λ) is the EL spectrum) [9] since the standard photometric curve V(λ) exhibits a peak of 555 nm and its intensity decreases rapidly from 555 nm to blue and red region [10].

 

Fig. 3. (Color online) The T(λ) and R(λ) characteristics of cathodes (a) and the R(λ) characteristic of the anode (b) of devices A, B, and C, and (c) resonant emission enhancement factors G_cav(λ) normalized to Z as described in the text.

Download Full Size | PDF

 

Fig. 4. (Color online) Dependence of EL spectra on the viewing angle of 0-80° for devices A (a), B (b), and C (c).

Download Full Size | PDF

Figure 4 shows the dependence of the EL spectra on the viewing angle. Figure 4(a) and (b) depicts the EL spectra of devices A and B with the angle changing from 0 to 80°. A large blueshift of the EL spectral peak of 28 and 36 nm occurs in devices A and B, with CIE coordinates change ranges of (±0.0123,±0.1421) and (±0.0717, ±0.0491), respectively. As a comparison, device C shows only 8 nm blueshift with CIE coordinates changes of (±0.0104, ±0.0284) after introducing a 97 nm-thick BCP film as the light outcoupling layer, which is mainly due to the effective reduction of microcavity effects by reducing cathode reflectivity with the BCP layer.

 

Fig. 5. (Color online) (a) Luminous reflectance of devices A, B, and C, and (b) reflection of Ag (23 and 12 nm), Ag (12 nm)/Sm (11 nm), and BCP (97 nm)/Ag (12 nm)/Sm (11 nm).

Download Full Size | PDF

The PCRs of devices A, B, and C are investigated with the formulas (2) [11] and (3) [12]:

and

where ere L_on and L_off are the brightness of a pixel at on- and off-state, respectively, L is the ambient illumination, R_L is the luminous reflectance, V(λ) is the standard photometric curve, R(λ) is the optical reflection of the device and S(λ) is the light source. Here, the light source is utilized with the spectrum of the blackbody radiation at 5777 K, which is similar with the spectrum of sun [13]. In the calculation, the brightness of the devices at on- and off-state is set to be 1000 and 0 cd/m². R_L for devices A, B, and C is calculated with the transfer matrix theory and depicted in Fig. 5(a). The PCR of device C is 88:1 under an ambient illumination of 140 lx, corresponding to an enhancement of 100 and 36% over those of devices A and B. The obvious improvement of PCR in device C originates from the reduction of R_L in the whole visible region. Here, the thickness of Ag in the SmAg cathode plays an important role in the anti-reflection of ambient light. Fig. 5(b) shows a simple comparison of the reflection of a 23 and 12 nm-thick Ag layer. A thin Ag layer effectively reduces the reflection of ambient light. However, not as expected, another Sm thin layer (11 nm) under Ag is not beneficial to the reduction of the cathode reflection. The BCP layer with a refractive index of ~1.6 and an extinction coefficient of ~0.1 inserted between air and Ag helps to get a lower reflection in the visible range since it can reduce the abrupt change of the refractive index and extinction coefficient during the light transmission.

4. Conclusions

In summary, we fabricate green TEOLEDs based on the Alq₃ emission with the Sm/Ag cathode and a comparative Al/Ag cathode. The device with the SmAg cathode demonstrates a lower brightness and luminous efficiency than those with the AlAg cathode due to a weak microcavity effect and a relative low electron injection because of a larger energy barrier of LiF/Sm than that of LiF/Al. The deposition of the BCP outcoupling layer onto the SmAg cathode improves not only device brightness and efficiency but also the viewing angle and the pixel contrast ratio. In addition to an enhancement factor of ~1.4 for brightness and efficiency, the TEOLED with a 97 nm-thick BCP layer exhibits only 8 nm blueshift of EL spectra with a little CIE coordinates change of (±0.0104, ±0.0284) and also shows an improved pixel contrast ratio of 88:1 under an ambient illumination of 140 lx and a device brightness of 1000 cd/m². The corresponding work mechanism has been investigated with the transfer matrix theory to explain these phenomena brought by the BCP outcoupling layer.