1. Introduction

Organic light emitting devices (OLEDs) have been subjects of intensive studies due to their various merits for flat-panel display applications, since Tang and VanSlyke reported the first practical bottom-emitting OLED, which consist of two organic layers sandwiched between a highly reflective Mg:Ag top cathode and a transparent indium tin oxide (ITO) bottom anode on glass substrates for light output [1]. For high performance full color display applications, active-matrix driven OLEDs are required to be integrated on electronic components such as Si based complementary metal-oxide semiconductor (CMOS) or thin-film-transistor (TFT) circuits. It is therefore desirable to develop the top-emitting OLEDs which emit light from the (semi-)transparent top electrodes for high display quality and large aperture ratios of pixels [2, 3]. A variety of top-emitting OLEDs fabricated on Si wafers have been reported for Si compatible electroluminescence (EL) and microdisplay applications [2–6]. However, the most reported devices that directly used either n- or p-type Si as the charge injection electrodes exhibited low efficiency, mainly due to the high absorption of Si in the visible light which greatly degrades the performance of the OLEDs [4–6]. For instance, the EL and power efficiencies ever reported for tris (8-hydroxyquinoline) aluminum (Alq₃) based OLEDs with p-type Si anodes were less than 0.8 cd/A and 0.2 lm/W [6]. In other works, the Si wafers were used only as the substrates, on which a highly reflective metal layer was formed and used as the charge injection electrode [2, 3]. The OLEDs, either conventional bottom-emitting or top-emitting, with the highly reflective back electrodes, exhibit rather strong reflection and low contrast under strong ambient illumination, which could become a serious problem for outdoor display applications [7]. Many approaches have been used to obtain the high-contrast conventional bottom-emitting OLEDs by employing the black cathodes based on the principle of destructive optical interference [8–10], or the light-absorbing layers [11, 12]. Recently, Yang et al. reported a high-contrast top-emitting OLED using the Mo bottom anode with a moderate reflectivity and the anti-reflection coating capped on the semitransparent top cathode [7]. While Lee et al. reported a high-contrast display by stacking a black reflective liquid crystal display and an OLED [13].

In this letter, we report high-efficiency and high-contrast phosphorescent top-emitting OLEDs employing the p-type Si bottom anodes with low reflectance and the Cs₂CO₃/Ag top cathodes with high transmittance for effective hole and electron injection. These devices, with the green electrophosphorescent material fac tris (2-phenylpyridine) iridium [Ir(ppy)₃] doped emitting layer, show peak external quantum and power efficiencies of 3.5% (12 cd/A) and 4.5 lm/W, which is, to the best of our knowledge, the highest value reported for OLEDs using Si wafers as electrodes. Moreover, these devices exhibit significantly higher contrast comparing to the conventional bottom-emitting and top-emitting OLEDs with the highly reflective back electrodes.

2. Experimental results and discussion

The structure of the top-emitting OLEDs with the p-type Si bottom anodes (device A) is shown in Fig. 1(a). The p-type Si (100) wafers used (from Fraunhofer’s Institute for Photonic Microsystems, Germany) were heavily doped (10¹⁹ cm^-3 boron) with a work function of 4.6 - 4.9 eV [4], a resistivity of 8×10^-3 - 2×10^-2 Ω cm, and a low reflectance of around 40% - 50% over most of the visible wavelength. We did not try to etch away the ~ 2 nm native oxide layer on top of the p-type Si wafers since this native oxide layer could enhance the hole injection due to the tunneling effect [4].

 

Fig. 1. Schematic illustration of the device structures.

Download Full Size | PDF

For comparison, the conventional top-emitting devices on Si wafer substrates [device B, shown in Fig. 1(b)], on which a 100-nm-thick Al layer was evaporated as the highly reflective bottom anode and a 2-nm-thick MoO₃ layer was used to improve hole injection [14], were fabricated. Both the device A and the device B employed the bilayer top cathodes, combining a 2-nm-thick Cs₂CO₃ layer for efficient electron injection and a 20-nm-thick semitransparent Ag layer [15], further capped with a 40-nm-thick Alq₃ layer for improving the transmittance of the top cathodes up to ~ 70% around 500 nm [16]. The conventional bottom-emitting devices [device C, shown in Fig. 1(c)] with the same organic multilayer structure sandwiched between the transparent ITO anodes and the highly reflective Cs₂CO₃ (2 nm)/Ag (100 nm) cathodes were also fabricated as the benchmark devices. In these devices, a 40-nm-thick 4, 4’-bis [N-(1-naphthyl-1-)-N-phenyl-amino]-biphenyl (NPB), a 20-nm-thick 4,4-N,N’-dicarbazole-biphenyl (CBP) doped with 4 wt % Ir(ppy)₃, a 10-nm-thick 2, 2’, 2”-(1, 3, 5-benzinetriyl) tris(1-phenyl-1-H-benzimidazole) (TPBi), and a 30-nm-thick Alq₃, were used as hole-transporting, light-emitting, hole-blocking, and electron-transporting layers, respectively. All material layers in the devices were thermally evaporated in sequence in a multi-source vacuum chamber at a base pressure of around 10^-6 Torr without breaking vacuum and the device characterization were performed in nitrogen glove box. The current density-voltage characteristics of the devices were measured by the Keithley 236 SMU. The forward direction photons emitted from the devices were detected by placing the calibrated Hamamatsu S1133 silicon photodiode very close onto the top of the devices. The luminance and external quantum efficiencies of the devices were inferred from the photocurrent of the photodiode.

Figure 2(a) compares the typical current density-luminance-voltage characteristics of the devices. It is obvious that the device A exhibits a substantially larger current density compared to the device B and the device C. For instance, at a driving voltage of 12 V, the device A generates a current density of 20 mA/cm², compared to 6.2 mA/cm² and 3.5 mA/cm² for the device B and the device C, respectively. These results clearly demonstrate that the hole injection from the p-type Si anodes to NPB is much stronger than those from the Al/MoO₃ and the ITO anodes, which is consistent with the previous works [5, 6], mainly due to the barrier reducing for hole injection at p-type Si/ NPB interface resulting from the alignment of the p-type Si Fermi level with the highest occupied molecular orbits of NPB induced by the ultra thin native oxide, and the efficient tunneling of holes [4]. The device B exhibits a higher current density than the device C which indicates that the MoO₃ modified Al anodes could effectively facilitate hole injection [14]. All devices studied here show similar luminance characteristics below 10 V, with a turn-on voltage of 3.5 V for 1 cd/m², and 7.5 V for obtaining 100 cd/m². However, the luminance of device A exhibits relatively fast increase at higher voltage maybe due to the more efficient tunneling of holes from p-type Si anodes to NPB through the ultra thin native oxide.

As shown in Fig. 2 (b), the peak external quantum and power efficiencies of the device C (the benchmark device) are around 11.5% (40 cd/A) and 13.5 lm/W, which are among the best values for the CBP:Ir(ppy)₃ OLEDs previously reported [17]. The peak external quantum and power efficiencies of device A are around 3.5% (12 cd/A) and 4.5 lm/W, approximately 30% of that of the device C. Such efficiency reduction is due to the low reflectance of the p-type Si bottom anodes. Another possible reason could be attributed to the strong hole injection from the p-type Si anodes, which leads to unbalanced charge recombination and reduced efficiency [5, 6]. Although these disadvantages, high-efficiency top-emitting OLEDs with p-type Si anodes were obtained in this study due to improved charge balance and out-coupling efficiency resulting from the efficient Cs₂CO₃/Ag electron injection cathodes with high transmittance [15, 16]. While the device B shows the peak external quantum and power efficiencies of around 12% (43 cd/A) and 38 lm/W at 1 cd/m² and remains 6% (20 cd/A) at 1000 cd/m², which are the typical values of phosphorescent top-emitting OLEDs previously reported by Riel et al [18]. It should be noted that we have not optimized the microcavity effects in terms of efficiency in this study. From comparison of the device A and the device B, it is evident that higher efficiency could be obtained by using a highly reflective bottom electrode (e.g. Al in our case), resulting in increased reflection of the devices. Thus there is trade off between the efficiency and contrast.

 

Fig. 2. (a) Current density-luminance-voltage and (b) EL efficiency characteristics of the devices.

Download Full Size | PDF

The measured reflectance of the device A is around 20% over most of the visible region, significantly lower than ~70% for the device B and ~90% for the device C, as shown in Fig. 3(a). The luminous reflectance (RL) to human eyes under a light source of D65 and contrast-ratios (CRs) of the devices were calculated according to the Ref. 9, the Ref. 11 and the Ref. 13 [9, 11, 13]. A RL of 23.9% for the device A was obtained, which is approximately 3 times and 4 times lower than 69.2% for the device B and 90.3% for the device C, respectively. The calculated CRs of the devices shown in Fig. 3(b) indicate greatly enhanced CR of the device A. For instance, the CR of the device A reaches 10.5, approximately 2.5 and 3 times higher than 4 for the device B and 3.5 for the device C, operating at 100 cd/m² under 140 lx of ambient lighting.

 

Fig. 3. (a) Measured reflectance spectra and (b) calculated contrast-ratios of the devices under 140 lx of ambient lighting.

Download Full Size | PDF

Photos in Fig. 4(a) show the “off” and “on (operating at 100 cd/m²)” status of the devices under 565 lx of strong white fluorescent light illumination. The “off” device A is very dark and the “on” device A looks pretty green with a CR of 3.5. While the “off” device B and C is quite bright due to the strong reflection of the ambient illumination. The “on” device B looks yellowish due to microcavity effects with a CR of 1.8. The “on” device C looks whitish due to the strong reflection of the ambient lighting with a CR of 1.5. As shown in Fig. 4(b), the EL spectra of the device A and the device C is identical and well agree with the result previously reported [17]. Owing to the microcavity effects, the corresponding normal-direction EL spectrum of the device B is modified and slightly shifted to longer wavelengths, which resulting in a color shift [18].

 

Fig. 4. (Color online) (a) Photos of the devices (with the active area of 6×6 m²) under 565 lx of white fluorescent light illumination. (b) EL spectra of the devices.

Download Full Size | PDF

3. Conclusion

In summary, we have demonstrated high-efficiency and high-contrast phosphorescent top-emitting OLEDs with p-type Si anodes. By employing the efficient Cs₂CO₃/Ag electron injection top cathode with high transmittance, we achieved peak external quantum and power efficiencies of 3.5% (12 cd/A) and 4.5 lm/W, which is the highest value reported for OLEDs using Si wafers as electrodes. Moreover, these devices exhibit significantly enhanced CRs due to the low reflectance of p-type Si bottom anodes. Our results suggest that the device structure may be promising for Si-based optoelectronics and microdisplay applications. The further optimization, for example, inserting an effective hole blocking layer to balance charge recombination for increasing the efficiency [19], using electrically doped charge transport layers for reducing the driving voltage is in progress [20].