Optical and Quantum Electronics (2006 38:1091 1099 Springer 2007 DOI 10.1007/s11082-006-9057-1 Optimization of microcavity OLED by varying the thickness of multi-layered mirror albert w. lu 1, j. chan 1, a.d. rakić 1, alan man ching ng 2,, a.b. djurišić 2 1 School of Information Technology and Electrical Engineering, The University of Queensland, Brisbane, Qld 4072, Australia 2 Department of Physics, The University of Hong Kong, Pokfulam Road, Hong Kong, Hong Kong ( author for correspondence: e-mail: alanalfa@gmail.com Received: 1 September 2006; Accepted: 1 December 2006 Abstract. We optimized the emission efficiency from a microcavity OLEDs consisting of widely used organic materials, N,N -di(naphthalene-1-yl-n,n -diphenylbenzidine (NPB as a hole transport layer and tris (8-hydroxyquinoline (Alq 3 as emitting and electron transporting layer. LiF/Al was used as a cathode, while metallic Ag was used as an anode material. A LiF/NPB bi-layer or NPB layer on top of the cathode was considered to alter the optical properties of the top mirror. The electroluminescence emission spectra, electric field distribution inside the device, carrier density, recombination rate and exciton density were calculated as a function of the position of the emission layer. The results show that for optimal capping layers thicknesses, light output is enhanced as a result of the increase in both the reflectance and transmittance of the top mirror. Once the optimum structure has been determined, the microcavity OLED devices were fabricated and characterized. The experimental results have been compared to the simulations and the influence of the thickness of the mirror layers, emission region width and position on the performance of microcavity OLEDs was discussed. Key words: microcavity OLEDs, optimization 1. Introduction OLEDs have emerged as a potential candidate for display applications due to their prominent advantages in size, brightness and wide viewing angle (Brutting et al. 2001; Hung and Chen 2002. Since the development of first prototype of OLED by Tang and VanSlyke, intense efforts had been put in to improving device performance (Tang and VanSlyke 1987. The optical microcavity with metallic mirrors is a simple and effective device architecture for light extraction and color purity improvement (Benisty et al. 1998; Neyts 1998; Neyts et al. 2000; Riel et al. 2003. Recently Riel et al. demonstrated that the addition of dielectric layers on top of the metallic cathode can lead to further improvements in optical intensity (Riel et al. 2003. In order to optimize the device, modeling of OLED characteristics is required to better understand the physical processes affecting the device performance. Device modeling based on electrical and optical models have
1092 A.W. LU ET AL. been documented separately in numerous studies, however, comprehensive device simulation that includes both electrical and optical models has been scarce (Ruhstaller et al. 2003; Lee et al. 2004. In this work, we aim to provide a detailed investigation into cavity design and thickness optimization of a bi-layer organic structure with multi-layered mirror by examining both electrical and optical characteristics. The paper outlines the optical and electrical simulation models used including the simulation parameters and simulated experimental conditions. The outcome of simulation is then compared with the experiment. 2. Simulation model 2.1. Electrical model The electrical transport inside the OLED is modeled by the one-dimensional time-independent drift-diffusion model (Blades and Walker 2000; Martin et al. 2002; Webster et al. 2004, which solves for a self-consistent solution of electron density, n, hole density, p and potential ψ implemented in the semiconductor solver Atlas (Silvaco 2000. The basic model includes: The continuity equation for n (electrons and p (holes d dx d dx ( μ n n dψ ( μ p p dψ dx + D dn n dx dx + D dp p dx = R = R, (1 where μ n and μ p are the electron and hole mobilities and D n and D p are diffusion constants and R is the recombination rate. The μ and D are frequently related by the well known Einstein relation. It was shown recently that, although for organic semiconductors the relationship between μ and D is dependent on the carrier concentration doping level and the LUMO and HOMO level of the host and dopant, the traditional Einstein relation still applies for low carrier concentration and is assumed to be valid in most analytic models (Peng et al. 2006. The carrier mobilities are modeled by the field-dependent form: [ ] E μ n (E = μ n0 exp E 0 [ ] E μ p (E = μ p0 exp, (2 E 0
OPTIMIZATION OF MICROCAVITY OLED 1093 with μ n0 and μ p0 are the zero field mobilities, E is the electric field and E 0 is the constant known as characteristic field. The recombination rate is taken to be optical only and modeled by the Langevin recombination coefficient γ (Martin et al. 2002; Webster et al. 2004: R opt = γ(pn n 2 i γ = 4πeμ R εε 0, (3 where n i is the intrinsic concentration and μ R is effective recombination mobility, taken to be the larger of the electron and hole mobilities in the material, εε 0 is the permittivity of the material. The Poisson s equation d 2 ψ dx 2 = e εε o [ p(x n(x + ND N A ], (4 where N D and N A are the ionized donor and acceptor dopant concentrations. These equations are solved for the p n junction structure using Schottky contact boundary conditions between a metal (which also serves as the reflecting surface for optical modeling and the organic layer at the anode and the cathode. The barrier heights governing carrier injections are: φ bn for electrons and φ bp for holes and are related to the metal work function φ m of the electrodes and the electron affinity of the organic material χ c : φ bn = (φ m χ c φ bp = (E g φ bn = E g (φ m χ c. (5 The continuity equations and the Poisson equation are solved to obtain the carrier concentrations, electric field distributions and recombination rate. The thickness of recombination region can be determined from the recombination rate, which can be used to estimate the width of emission region (taking into account of exciton diffusion to be included into the optical model. 2.2. Optical model Light output was calculated using the method based on the equivalence between the probability of photon emission and the power radiated by a classical dipole antenna (Neyts 1998; Neyts et al. 2000; Riel et al. 2003. First the OLED is treated as a microcavity with an emissive layer (Alq 3 in
1094 A.W. LU ET AL. this case sandwiched between two mirrors (consisting of thin film layers. The total power F emitted by a dipole antenna relative to the power output of the same dipole in an infinite medium (so in an infinite medium, F equals to one is given by: F = 0 K (κ dκ 2, (6 where K is the power density per unit dκ 2. The power density (K can be resolved into the TM and TE component, with each component separated into power densities for dipoles oriented parallel and perpendicular to the z-axis. With this in mind, the power densities can be defined as: K TM = 3 4 R [ κ 2 k 2 e k z,e ( 1 a TM + ( 1 a TM 1 a TM ], (7 K TE = 0, (8 [ K TM = 3 ( ( 8 R k z,e 1 + a TM + 1 + a TM ke 3 1 a TM ] (9 [ = 3 8 R 1 k e k z,e K TE ( 1 + a TE + ( 1 + a TE 1 a TE ] (10 where e denotes the emissive layer (sandwiched between top and bottom mirrors and a is the reflection coefficient of the mirror with respect to the location of the dipole, defined as: a TM/TE +/ = r TM/TE e,+/ exp ( 2jk z,e z +/ a TM/TE = a TM/TE + a TM/TE (11 where z + is the dipoles distance from the top mirror, z is the dipoles distance from the bottom mirror and r is the amplitude reflection coefficient of the top and bottom mirrors calculated using the modified transfer matrix approach of Katsidis and Siapkas (Mitsas and Siapkas 1995; Katsidis and Siapkas 2002.
OPTIMIZATION OF MICROCAVITY OLED 1095 Elec tric Fiel d ( V/cm 2x10 6 1.8x10 6 1.6x10 6 1.4x10 6 1.2x10 6 10 6 8x10 5 6x10 5 E-field 10 19 10 14 10-1 10-6 10-11 4x10 5 10-16 0.05 0.06 0.07 0.08 0.09 0.10 a NPB Alq 3 Electron Hole Distance along the device (µm 10 9 10 4 Ca rrie r D e nit s y ( c m - 3 Optical f ied l in t ens i t y (a r b. u ni t s 1.0 0.8 0.6 0.4 0.2 0.0 Ag NPB Alq 3 LiF/Al 0.00 0.05 0.10 0.15 0.20 0.25 b Distance along the device (µm Fig. 1. (a Simulated carrier density and electric field for the device. (b Internal optical field distribution. 3. Experiment details The materials used (Alq 3 and NPB from H. W. Sands Corp. were purified by vacuum sublimation before fabrication. Devices were fabricated by vacuum deposition. The pressure during evaporation was of the order 10 4 Pa. The evaporation rate was kept at 1 Å/s. The distance from source to film was about 23 cm to ensure uniformity of film thickness, and the substrate holder was rotating. The thickness of the films was controlled by a quartz thickness monitor. The electrodes consisted of Al and LiF films. The electroluminescence (EL spectra were measured using Control Development fiber optic spectrometer (PDA 512 USB, while devices were biased using a Keithley 2400 sourcemeter. 4. Results and discussion Firstly, general electrical properties of OLED structure were investigated. Figure 1(a shows the injection characteristics of the device. Optimum thickness of the cavity is determined using cavity equation, the thickness is determined to be approximately 120 nm with cavity order of 0. The optical field intensity distribution was then calculated for the device in order to align the antinode of the field with the light emission region of the device. This can be seen in Fig. 1(b The simulation shows that the antinode is located at approximately the middle of the cavity. To align the center of recombination layer with this antinode, we used a device with 60 nm NPB and 60 nm Alq 3 (as the recombination layer centers around the interface of the HTL/ETL layers, see Fig. 2(a and (b.
1096 A.W. LU ET AL. Fig. 2. (a Recombination rate (b Exciton density. (c Reflectance and Transmittance versus Al thickness. The width of recombination region obtained from the above simulations can be used in the optical model to calculate the electroluminescent intensity (i.e. optical intensity. Figure 2(c shows the reflectance and transmittance plot for the unclad LiF/Al top mirror as a function of film thickness, calculations were performed for incidence from Alq 3. To obtain the beneficial effect of the optical microcavity device has to have mirrors with considerable reflectivity combined with minimal loss in the mirror structure and sufficient transmittance to allow efficient light extraction. Taking these factors into consideration, we chose 20 nm as our aluminum layer film thickness. Figure 3 shows the reflectance and the transmittance of the LiF/NPB clad LiF/Al top mirror as a function of LiF and NPB thicknesses d 1 and d 2, respectively; Al film thickness was fixed at 20 nm. The general shape of the R(d 1, d 2 surface remains the same for a range of different Al thicknesses; the magnitude of the energy coefficients R and T is, however, dependent on the thickness of the Al film. It can be seen that for certain LiF and NPB thicknesses, we can obtain both a higher R and a higher T compared to that of the bare aluminum by minimizing the absorption within the metallic film. This claim is further supported up by the contour plot of peak optical intensity with respect to both LiF and NPB thickness shown in Fig. 4. It can be seen from Fig. 4 that device with optimum performance occurs in the two regions labeled C1 and C2. The results of the optimization process show that LiF, NPB thicknesses of 190 and 75-nm, respectively yield the maximum external optical intensity. Figure 7(a shows the simulated electroluminance spectra for the optimized device and the device without any dielectric cap. It can be seen that
OPTIMIZATION OF MICROCAVITY OLED 1097 Fig. 3. Reflectance (top graph and Transmittance (bottom graph versus LiF and NPB thickness (in nm. NPB thicknes s 10 1 n m 16 14 12 10 8 6 C1 0.047 0.065 0.056 0.065 0.056 0.047 0.074 0.065 0.0370.047 0.084 0.056 C2 4 0.074 0.084 2 4 6 8 10 12 14 16 18 20 LiF thickness 10 1 nm Fig. 4. Contour plot of peak optical intensity versus thickness. based on our model the optimized device shows a twofold improvement in the peak optical intensity. The intensity improvement can be attributed to both the increase in top mirror reflectance and transmittance and the reduction in absorption, as shown in Figs. 5 and 6. From Figs. 5 and 6, it can be seen that for the optimized device with a single layer cap (corresponding to region C1 in Fig. 4 and optimized device with two caps (region C2 have higher R, T and lower A within the wavelength of interest, as opposed to the device without capping, where the mirror characteristics (R, T and A remains more flat over the entire visible range. Thus the optimized capped devices will also display improved color purity compared to uncapped devices. These claims are vindicated by measurement on fabricated devices shown in Fig. 7(b is clear that the device with a dielectric
1098 A.W. LU ET AL. Refle ct ance (% 0.80 0.78 0.76 0.74 0.72 0.70 x20y190z75 x20y0z70 L=LiF x=al y=lif z=npb L0.5x20y0z0 L0.5x20y0z70 L0.5x20y190z75 x20y0z0 Tr an smittan ce (% 0.09 0.08 0.07 0.06 0.05 0.04 0.03 x20y0z0 x20y190z75 L=LiF x=al y=lif z=npb L0.5x20y0z0 L0.5x20y0z70 L0.5x20y190z75 x20y0z70 0.68 0.02 0.66 300 400 500 600 700 800 900 Wavelength (nm 0.01 300 400 500 600 700 800 900 Wavelength (nm Fig. 5. Reflectance (R and Transmittance (T spectrum for devices with different caps (film thickness in nanometers. 0.86 0.84 0.82 L=LiF x=al y=lif z=npb L0.5x20y0z0 L0.5x20y0z70 L0.5x0y190z75 0.80 R + T 0.78 0.76 0.74 0.72 0.70 300 400 500 600 700 800 900 Wavelength (nm Fig. 6. Sum of Reflectance and Transmittance (1-A Spectrum for devices with different caps (film thicknesses in nanometers. cap does indeed improve both the intensity and color purity (via reduced spectral width of the device, as predicted by our calculations. The differences between the simulated and measured results can be attributed to a number of factors related to the growth process and parameters of materials used. The uncertainty in the measured film thicknesses translates to the uncertainty of the resonant peak position. Also, the uncertainty incurred in extracting the optical constants form the ellipsometric measurements will also translate to uncertainty in the peak position and the spectral width. Regardless of this, experimental results demonstrate clearly the increase in peak intensity and spectral narrowing due to the microcavity effect predicted by the model.
OPTIMIZATION OF MICROCAVITY OLED 1099 ni s i y ( arb. u t Nor mali zed Inte nst 2.0 1.5 1.0 0.5 L=LiF x=al y=lif z=npb L0.5x20y190z75 L0.5x20y0z0 Opt ical In ten sity (a rb. uni ts 3.0 2.5 2.0 1.5 1.0 0.5 without cap -> LiF0.5Al20 with cap -> LiF0.5Al20NPB70 0.0 400 500 600 700 (a Wavelength (nm 0.0 400 500 600 700 b Wavelength (nm Fig. 7. (a Simulated electroluminance spectra of the devices with and without the dielectric caps (b Measured EL spectra of devices with and without dielectric cap. 5. Conclusion We have demonstrated that by using a relatively simple structure comprising one- or two-layer dielectric caps on top emitting Organic Light Emitting Diodes, we can simultaneously increase the reflectance and transmittance and reduce the absorption of the top mirror, thus boosting the output optical intensity and reducing the spectral width of the device. References Blades C.D.J. and A.B. Walker. Proceedings of Synthetic Metal 2nd International Conference on Electroluminescence of Molecular Materials and Related Phenomena May 15-May 18 1999, Vol 111, p. 335, 2000. Brutting, W., S. Berleb and A. G. Muckl. Organic Electron.: phys. mater. Appl. 2 1, 2001. Benisty, H., H. Neve and C. Weisbuch. IEEE J. Quantum Electron. 34 1612, 1998. Hung, L.S. and C.H. Chen. 39 80, 2002. Katsidis C.C. and D.I. Siapkas. Appl. Opt. 41 3978, 2002. Lee, C.-C., M.-Y. Chang, Y.-D. Jong, T.-W. Huang, C.-S. Chu and Y. Chang. Jpn J. Appl. Phys. Part 1: Reg. Pap. Short Notes Rev. Pap. 43 7560, 2004. Mitsas C.L. and D.I. Siapkas. Appl. Opt. 34 1678, 1995. Martin, S.J., G.L.B. Verschoor, M.A. Webster and A.B. Walker, Organic Electron. 3 129, 2002. Neyts, K.A. J. Optical Soc. Am. A: Optics Image Sci. Vis. 15 962, 1998. Neyts, K., P. De Visschere, D.K. Fork and G.B. Anderson. J. Optical Soc. Am. B (Optical Phys. 17 114, 2000. Peng, Y.Q., J.H. Yang and F.P. Lu. Appl. Phys. A, 83 305, 2006. Riel, H., S. Karg, T. Beierlein and W. Rielβ. J. Appl. Phys. 94 5290, 2003. Ruhstaller, B., T. Beierlein, H. Riel, S. Karg, J.C. Scott and W. Riess. IEEE J. Select. Top. Quant. Electron. 9 723, 2003. Silvaco, ATLAS User Manual, Santa Clara: Silvaco International, 1, 2000. Tang, C.W. and S.A. VanSlyke. Appl. Phys. Lett. 51 913, 1987. Webster, M.A., J. Auld, S.J. Martin and A.B. Walker. Proc. SPIE 5214 300, 2004.