Open Access
Add to BibSonomyAbstract
The light extraction efficiency of OLEDs with a nano-sized random scattering layer (RSL-OLEDs) was analyzed using the Finite Difference Time Domain (FDTD) method. In contrast to periodic diffraction patterns, the presence of an RSL suppresses the spectral shift with respect to the viewing angle. For FDTD simulation of RSL-OLEDs, a planar light source with a certain spatial and temporal coherence was incorporated, and the light extraction efficiency with respect to the fill factor of the RSL and the absorption coefficient of the material was investigated. The design results were compared to the experimental results of the RSL-OLEDs in order to confirm the usefulness of FDTD in predicting experimental results. According to our FDTD simulations, the light confined within the ITO-organic waveguide was quickly absorbed, and the absorption coefficients of ITO and RSL materials should be reduced in order to obtain significant improvement in the external quantum efficiency (EQE). When the extinction coefficient of ITO was 0.01, the EQE in the RSL-OLED was simulated to be enhanced by a factor of 1.8.
© 2014 Optical Society of America
1. Introduction
Organic light emitting diodes (OLEDs) have received significant attention recently with regards to mobile and large scale displays as well as lighting applications. The unique thin film light emitting structure of these devices consists of several organic materials that enable flexible and transparent self-emitting displays that are not achievable with other technologies. In order to improve the performance of the OLED device in addition to the development of novel organic materials with superior quantum efficiency, the device structure should be optimized for better optical outcoupling efficiency. In an ordinary OLED, half of the light generated from excitons in organic layer is captured directly inside the organic-ITO waveguide followed by further losses when propagating via total internal reflection on a glass-air interface. As a result, only 20% of the light generated is radiated out of the device. By designing an optical structure to enhance the optical output it is therefore possible to improve the power efficiency as well as the lifetime [1, 2].
As the most of the light is captured in the ITO-organic waveguide, a scattering element located close to the waveguide is the most effective method to increase the optical output efficiency. A light scattering photonic crystal structure was placed under the ITO electrode along with a high refractive index layer [3]. To improve the scattering efficiency this optical scattering structure was inserted in the middle of the ITO and organic layer in which a strong optical field intensity occurs [4, 5]. To overcome the spectral shift and angular dependence of the radiation pattern caused by periodic scattering elements, a randomly distributed nano-pattern was fabricated by simple wet etching of the ITO layer [6], and a randomly dispersed nanopillar array was formed by angled deposition of a dielectric material [7]. The transfer of a buckling structure to the entire device layer was also effective at enhancing the radiation outcoupling of the guided light [8].
To reduce the burden of repetitive experiments, simulation of the proposed optical structure before the fabrication is important. Optical modeling using the finite difference in time domain (FDTD) method has been widely used for the optical analysis of OLEDs composed of multiple submicron thick layers and various nano-patterns [9]. For the OLED with the photonic crystal structure, FDTD simulations have been carried out in order to estimate the effect of the photonic crystal dimension [10, 11]. Furthermore, the influence of surface plasmon absorption caused by a metal layer in the OLED was analyzed by the FDTD simulation [12]. It has also been applied to the design of optical outcoupling in inorganic LED devices [13].
In this work, FDTD simulations were used to investigate OLEDs containing a random scattering layer (RSL). The RSL is a pattern consisting of nano structures with random period and size, and was chosen over a periodic scattering pattern as it has the potential to overcome the spectral shift and angular dependence of the output light and thus be adopted in commercial OLEDs. For the simulation of the RSL-OLEDs, a planar light source with no spatial coherence was incorporated into the model, and the temporal coherence of the Gaussian pulse was adjusted to have the same spectral width as the measured spectrum. Throughout the simulation, it was found that the outcoupling efficiency was highly dependent on the optical absorption of the ITO, metal, and organic materials. The simulations were then compared to the experimental results of fabricated RSL-OLEDs, where it was confirmed that the FDTD simulation is useful for estimation of the effect of the proposed RSL.
2. Design of the RSL-OLEDs and the FDTD simulation conditions
A typical OLED device consists of an ITO anode, a hole transporting layer (HTL), a light emitting layer (EML) an electron transporting layer (ETL), and a metal cathode. As the HTL, EML and ETL have similar optical refractive indices they can be regarded as a single layer in optical simulations. In the RSL-OLEDs, the scattering layer (composed of glass or silica) and a high index buffer layer (HIBL) are located under the ITO anode, as shown in Fig. 1. The HIBL is deposited onto the random pattern in order to form a planar surface. As the optical refractive index of the HIBL is close to that of the ITO layer, light will naturally penetrate into the HIBL and then be scattered by the RSL. Compared to periodic patterns that diffract guided light, the RSL causes scattering with no angular dependence.
Fig. 1 OLED device structure consisting of an Al cathode, organic layers, an ITO anode, a high index buffer layer, and a random scattering layer.
For the simulation of the RSL, a planar light source is preferable over a single point light source because the point source gives significantly different results depending on the overlap of the random pattern with the position of the point source. An array of point sources, with randomly set initial phases on each point, as shown in Fig. 2(a), were used to introduce a spatially incoherent planar light source. With this incoherent planar source, the optical radiation pattern was changing for each time of calculation as shown in Fig. 2(a). The radiation flux measured by the sensor also oscillated, as shown in Fig. 2(b), hence the calculation should be repeated until the saturated average flux is obtained.
Fig. 2 (a) Electric field radiation patterns emitted from the dipole array with a randomly set initial phase on each dipole, and (b) output flux detected on the sensor that changed for each calculation, and the convergence of the average value as the number of calculations was increased.
Spectral broadening of the OLED is principally caused by an inhomogeneous broadening of the emitting molecules. Thus the temporal coherence of the light source has to be adjusted to reflect the actual spectral width of the OLEDs. For simplicity in the simulation, however, each dipole of the planar source was adjusted to produce homogeneous broadening corresponding to the measured spectrum. A Gaussian pulse of duration 18.8 fs resulted in a spectral bandwidth of 70 nm, as shown in Fig. 3.
Fig. 3 (a) Gaussian pulse wave packet used in the simulation (pulse duration = 18.8 fs), and (b) the resultant spectrum of the light source compared to the experimental measurement.
When there is a scattering layer designed to extract the guided light (in OLEDs), even though the dipole source emits a pulse, the output light will last much longer than the original pulse duration, and will eventually stop after all the light has been extracted from the device. In previous work, therefore, the outcoupling efficiency was dependent on the flux integration time [10]. In a practical sense, however, even if there is a scattering layer, most of the light will be absorbed and thus still cannot escape from the device. In the simulations we accounted for the practical absorption coefficient of each material in the OLEDs, as summarized in Table 1. The highest absorption is caused by the ITO layer and is largely dependent on the deposition conditions of the ITO compound. (A typical 100-nm thick ITO film absorbs 5% in vertical transmission and has an extinction coefficient ni of 0.02.) Moreover, Al has extremely high absorption of TM polarized light through surface plasmon absorption. The effect of the extinction coefficient is shown in Fig. 4. When there was no absorption, the accumulated flux increased continuously, whereas it saturated when the ITO possessed a finite extinction coefficient (ni), of 0.02 and 0.04. When ni was 0.02, the output flux was saturated after about 1000 fs.
Table 1. Summary of the material parameters used for the FDTD simulationsa
Fig. 4 Accumulated output flux of the RSL-OLED with respect to the extinction coefficient of the ITO and Al. With no absorption the flux is continuously increasing, whilst it saturates when the practical absorption of ITO is reflected in the simulation.
The absorption coefficient of organic material is smaller than that of ITO due to the Stoke’s shift. However, when an organic thin film is fabricated to be in contact with another material, the organic material could result in higher absorption coefficients by forming an interfacial energy states [14]. To reflect that effect, it may be necessary to measure the complex refractive index of multilayer organics evaporated onto an ITO substrate. The HIBL material was prepared by dispersing metal oxide nano-particles within a transparent polymer. Initially the HIBL had an order of magnitude lower absorption than the ITO, however, the loss could be significantly increased by aggregation of the nanoparticles and yellowing of the polymer. All of these absorption mechanisms should be reflected in the FDTD simulation in order to increase the accuracy of the results.
3. Light extraction efficiency calculated by FDTD simulations
Considering the total calculation time and required memory, for simulations of an OLED device with multi-layer films of the order of tens of nanometers in thickness the spatial grid (Δx, Δy, Δz) was set to 20 nm in each direction. For a calculation window of 6 x 6 x 6 μm3, this results in a total of 27 million calculation points. As shown in Fig. 5, the effective calculation boundary was extended in the lateral y and z directions by utilizing the mirror effect at the reflection boundary. The light was extracted along the x direction, and a planar power detector was located underneath the glass substrate. In the RSL-OLED of Fig. 5(b), the HIBL was located underneath the ITO anode layer, and the RSL was located above the glass substrate. The dipole source array was located under the Al cathode at a distance of d. The RSL pattern was generated using a random function; the resultant pattern closely resembled that of the actual fabricated sample. In this study, the FDTD program available from PhotonDesign Co. (OMNISIMM) was utilized.
Prior to the outcoupling calculation, the effect of limited coherence of the light source on the micro cavity was investigated. In a typical OLED the cavity is formed between the Al cathode and the ITO-glass interface, however, in the RSL-OLED (with HIBL) the ITO-glass interface is replaced with a HIBL-glass interface. The glass-air interface is too far from the other interface to form an effective cavity. To observe the reduced micro-cavity effect due to the HIBL layer, the outcoupled flux was calculated as a function of the HIBL thickness and various source conditions. As shown in Fig. 6, when a spatially coherent CW source (i.e., a uniform plane wave) is present one can see large fluctuations of the flux as a function of the HIBL thickness. The presence of periodic oscillatory patterns strongly indicates the presence of micro-cavity. However, when the spatial coherence was removed, or the temporal coherence was reduced, the oscillation amplitude was significantly reduced, thus indicating that the micro-cavity effect had become negligible. For the case of a coherence time of 18.8 fs (coherence length of 5.6 μm), the micro-cavity effect disappeared if the HIBL thickness exceeded 300 nm. When the source was spatially incoherent, the emitting light diverged in a wide angle through the HIBL so that it became difficult to produce considerable interference after a round-trip in the cavity.
Fig. 6 Influence of the temporal and spatial coherence on the micro-cavity with a 280 nm thick organic layer. The spatially coherent CW light exhibits sinusoidal fluctuation of the flux as a result of the micro-cavity effect, whilst the temporally or spatially incoherent light exhibits negligible fluctuations, indicating that the micro-cavity effect has reduced.
The outcoupling efficiency or the external quantum efficiency (EQE) of an OLED is strongly dependent on the distance of the emitting source from the Al mirror, d, as well as the total thickness of the organic materials, to. The EQE of the reference OLED and RSL-OLED were therefore calculated as a function of these two variables, as shown in the contour graph of Fig. 7. The RSL-OLED had a scattering pattern with a spatial distribution of 200 - 500 nm, a height of 400 nm, a fill factor of 54%, and a HIBL thickness of 800 nm. The EQE values were calculated by determining the ratio of the flux extracted from the OLED device to the flux generated inside the device. By comparing these two results, one can observe a periodicity with respect to to in the reference device that was not present in the RSL-OLED due to disappearance of the micro cavity effect, as explained previously. In the reference device, the maximum EQE for d = 60 nm was 33.9 and 31.4% for to = 120 and 280 nm, respectively. The value of d was shorter than λ/4n = 75 nm because Al metal introduces an additional phase change over 180° [15]. For the RSL-OLED, as shown in Fig. 7(b), the EQE depended only on d, whilst the to dependence was greatly reduced due to the negligible micro-cavity effect. When d = 60 nm and to = 120, 180, 320 nm, the EQE approached 45.5%: the EQE enhancement by addition of an RSL was 34%.
Fig. 7 External quantum efficiency calculated for (a) the reference device which exhibited a periodic response due to the micro-cavity effect, and (b) the RSL-OLED which exhibited negligible micro-cavity effects.
The enhancement ratio was also calculated as a function of the refractive index of the HIBL layer for various fill factors of the RSL. With the two variables d and to fixed to provide the highest EQE in the reference device (d = 60 nm, to = 280 nm), the FDTD simulation was used to find the enhancement ratio, as shown in Fig. 8. As the refractive index of the HIBL increased, the enhancement ratio gradually increased due to the reduced reflectivity at the interface between ITO anode and HIBL was and the enhanced scattering by the higher index contrast between HIBL and RSL. For a fill factor of 54%, the highest enhancement ratio was obtainable because the greatest scattering occurred when a fill factor is closed to 50%.
Fig. 8 Enhancement ratio of the EQE when incorporating a RSL, as a function of the refractive index of the HIBL for various fill factors.
4. Comparison to experimental results
To confirm the usefulness of the FDTD simulations, the simulation results were compared to experimental results from our group. The structure of the fabricated OLEDs and the fabrication process of the RSL are described in detail in a previous publication by the authors [16]. In the fabricated device the ETL thickness was 70 nm and the emission material layer (EML) thickness was 10 nm. The total organic layer thickness was 240, 280, and 310 nm in the 3 samples studied and was varied by controlling the HTL thickness. The vertical transmission loss of the ITO thin film was determined to be 10%, and thus ni and α of the ITO thin film became 0.046 and 10400 cm−1, respectively. The nano-structure of the fabricated RSL prepared by the silver dewetting and dry etching process [17], and the HIBL layer formed onto the RSL are shown in Fig. 9. The size distribution of the nano-structures in the RSL was 200 - 500 nm, and the fill factor was around 31%. A metal-oxide dispersed polymer was spin-coated over the RSL to form a 800-nm thick HIBL. Surface planarization was successful, however, a slight air gap appeared between the RSL and the HIBL that may have reduced the effective index of the HIBL and thus reduced the scattering efficiency.
Fig. 9 (a) SEM image of RSL structure composed of SiO2 fabricated by the silver dewetting and dry etching process, and (b) SEM image of the good flatness obtained by the HIBL covering of the RSL.
The air gap has no effect on the carrier injection because it appears between glass substrate and ITO anode.
To compare the fabricated OLED to the simulation results (where d = 80 nm and ni = 0.046 for the ITO), the FDTD simulation was performed for the RSL-OLED with a fill factor of ~27%. The EQE and enhancement ratio of the fabricated device is compared to the simulations in Fig. 10. In the simulation results given in Fig. 10(a), the EQE of the reference device was dependent on the organic thickness, and was a maximum for to = 280 nm (where the micro-cavity effect was maximized). In contrast, the EQE of the RSL-OLED was independent of the organic thickness. In the experimental results, a similar behavior was observed with respect to the organic thickness dependence. The enhancement factor should be compared to the highest value obtained for the reference OLED. This occurred at an organic thickness of 280 nm, at which point the experimental result exhibited an enhancement of 1.24, whilst the FDTD result indicated an enhancement of 1.37. The discrepancy of these values could be explained by several reasons such as a reduced effective index of the HIBL due to the air gap shown in SEM image of Fig. 9(b); a higher absorption loss of the ITO; and increased absorption of the HIBL due to thermal yellowing.
Fig. 10 Comparison of (a) the simulation results (fill factor = 27%) to (b) the results of the fabricated device (fill factor = 31%) with d = 80 nm, and ni = 0.046.
As the result of Fig. 4 clearly indicated, the absorption of the ITO layer is a dominant factor when determining the output flux and the enhancement ratio. If the light penetrates the ITO vertically, as was the case for the reference OLED, absorption is not that important. However, light captured in the ITO-organic waveguide is significantly attenuated by the ITO absorption before it is extracted. To draw a final conclusion, we observed the effect of ITO absorption on the enhancement ratio. As shown in Fig. 11, as the extinction ratio of the ITO was reduced, the EQE of the RSL-OLED gradually increased whilst the reference OLED exhibited negligible change. Therefore, the enhancement factor also increased quite significantly, approaching a factor of 1.8 if the ITO extinction ratio is reduced to 0.01. The absorption coefficient of ITO can be largely affected by the deposition conditions [18, 19], and it is important to produce low-loss ITO and HIBL in order to maximize the EQE of RSL-OLEDs.
Fig. 11 The effect of the extinction coefficient of the ITO on the EQE and enhancement ratio of RSL-OLEDs. These results indicate that the EQE could be increased by a factor of 1.8 if the ITO has low extinction coefficient.
5. Conclusion
In the context of low energy consumption, device life time and enhanced luminance, improved efficiency of OLED is highly desirable. Random scattering layer may be used as a light extraction structure in OLED applications. In this work, we have investigated the light extraction performance of RSL using FDTD simulations. In order to properly reflect the light characteristics of OLED, we have constructed dipoles arrays to simulate a planar light source with a random spatial coherence and a certain temporal coherence. Due to the limited coherence, the device with a RSL and HIBL exhibited negligible micro-cavity effects, as confirmed by the experimental result of a RSL-OLED. The enhancement factors observed in the experimental results were close to that of the simulation results. It has been found that the EQE enhancement was strongly affected by the material absorption because the ITO-organic guided light was quickly absorbed before it was coupled out of the device. The extinction coefficient of ITO should be reduced to less than 0.01 in order to achieve an EQE enhancement of over 1.8 when incorporating a RSL.
Acknowledgments
This work was partially supported by a Korea Science and Engineering Foundation (KOSEF) grant (2012-001697) from the Ministry of Education, Science, and Technology, Korea, and ETRI Internal Research Fund from Ministry of Science, ICT and Future Planning. The authors would like to appreciate LG display for their partial support.
References and links
1. M.-H. Lu and J. C. Sturm, “Optimization of external coupling and light emission in organic light-emitting devices: modeling and experiment,” J. Appl. Phys. 91(2), 595–604 (2002).
2. V. Bulović, V. B. Khalfin, G. Gu, P. E. Burrows, D. Z. Garbuzov, and S. R. Forrest, “Weak microcavity effect in organic light-emitting devices,” Phys. Rev. B 58(7), 3730–3740 (1998). [CrossRef]
3. Y.-C. Kim, S.-H. Cho, Y.-W. Song, Y.-J. Lee, Y.-H. Lee, and Y.-R. Do, “Planarized SiNx/spin-on-glass photonic crystal organic light-emitting diodes,” Appl. Phys. Lett. 89(17), 173502 (2006). [CrossRef]
4. Y. Sun and S. R. Forrest, “Enhanced light out-coupling of organic light-emitting devices using embedded low-index grids,” Nat. Photonics 2(8), 483–487 (2008). [CrossRef]
5. J.-H. Jang, M.-C. Oh, T.-H. Yoon, and J.-C. Kim, “Polymer grating imbedded organic light emitting diodes with improved out-coupling efficiency,” Appl. Phys. Lett. 97(12), 123302 (2010). [CrossRef]
6. J.-H. Jang and M.-C. Oh, “Outcoupling enhancement of OLED with randomly distributed ITO pattern fabricated by maskless wet etching method,” J. Disp. Technol. 9(11), 900–903 (2013). [CrossRef]
7. C. H. Lee and J.-J. Kim, “Enhanced light out-coupling of OLEDs with low haze by inserting randomly dispersed nanopillar arrays formed by lateral phase separation of polymer blends,” Small 9(22), 3858–3863 (2013). [CrossRef] [PubMed]
8. W.-H. Koo, S.-M. Jeong, F. Araoka, K. Ishikawa, S. Nishimura, T. Toyooka, and H. Takezoe, “Light extraction from organic light-emitting diodes enhanced by spontaneously formed buckles,” Nat. Photonics 4(4), 222–226 (2010). [CrossRef]
9. A. Chutinan, K. Ishihara, T. Asano, M. Fujita, and S. Noda, “Theoretical analysis on light-extraction efficiency of organic light-emitting diodes using FDTD and mode-expansion methods,” Org. Electron. 6(1), 3–9 (2005). [CrossRef]
10. Y.-R. Do, Y.-C. Kim, Y.-W. Song, C.-O. Cho, H.-S. Jeon, Y.-J. Lee, S.-H. Kim, and Y.-H. Lee, “Enhanced light extraction from organic light-emitting diodes with 2D SiO2/SiNx photonic crystals,” Adv. Mater. 15(14), 1214–1218 (2003). [CrossRef]
11. A. O. Altun, S.-H. Jeon, J.-Y. Shim, J.-H. Jeong, D.-G. Choi, K.-D. Kim, J.-H. Choi, S.-W. Lee, E.-S. Lee, H.-D. Park, J. R. Youn, J.-J. Kim, Y.-H. Lee, and J.-W. Kang, “Corrugated organic light emitting diodes for enhanced light extraction,” Org. Electron. 11(5), 711–716 (2010). [CrossRef]
12. X.-L. Zhang, J. Feng, J.-F. Song, X.-B. Li, and H.-B. Sun, “Grating amplitude effect on electroluminescence enhancement of corrugated organic light-emitting devices,” Opt. Lett. 36(19), 3915–3917 (2011). [CrossRef] [PubMed]
13. W. J. Choi, Q.-H. Park, D.-H. Kim, H.-S. Jeon, C.-S. Sone, and Y.-J. Park, “FDTD simulation for light extraction in a GaN-based LED,” J. Korean Phys. Soc. 49(3), 877–880 (2006).
14. A. Aziz and K. L. Narasimhan, “Subband gap optical absorption and defects in Tris(8 hydroxy quinolato) aluminium,” Synth. Met. 131(1–3), 71–77 (2002). [CrossRef]
15. J. M. Bennett, “Precise Method for Measuring the Absolute Phase Change on Reflection,” J. Opt. Soc. Am. 54(5), 612–624 (1964). [CrossRef]
16. D.-H. Cho, J.-W. Shin, S.-K. Park, J.-H. Moon, B.-J. Lee, N.-S. Cho, J.-H. Han, J.-W. Huh, C.-W. Joo, J.-H. Hwang, H.-Y. Chu, and J.-I. Lee, “Light extraction for white OLEDs by nano-structures and high refractive index layer,” in Digest of Technical Papers of the 12th International Meeting on Information Display (IMID, Daegu, Korea, 2012), 182–183.
17. C.-I. Yeo, Y.-M. Song, S.-J. Jang, and Y.-T. Lee, “Wafer-scale broadband antireflective silicon fabricated by metal-assisted chemical etching using spin-coating Ag ink,” Opt. Express 19(S5Suppl 5), A1109–A1116 (2011). [CrossRef] [PubMed]
18. Y. Hu, X. Diao, C. Wang, W. Hao, and T. Wang, “Effects of heat treatment on properties of ITO films prepared by rf magnetron sputtering,” Vacuum 75(2), 183–188 (2004). [CrossRef]
19. Y.-S. Jung, Y.-W. Choi, H.-C. Lee, and D.-W. Lee, “Effects of thermal treatment on the electrical and optical properties of silver-based indium tin oxide/metal/indium tin oxide structures,” Thin Solid Films 440(1–2), 278–284 (2003). [CrossRef]
