Design of microcavity organic light emitting diodes with optimized electrical and optical...

Design of microcavity organic light emitting diodeswith optimized electrical and optical performance

Albert W. Lu and Aleksandar D. Rakić*School of Information Technology and Electrical Engineering, The University of Queensland, Brisbane,

QLD 4072, Australia

*Corresponding author: [email protected]

Received 12 March 2009; accepted 17 March 2009;posted 26 March 2009 (Doc. ID 108591); published 14 April 2009

A multivariable and multiobjective organic light emitting diode (OLED) design and optimizationprocedure is presented that produces a microcavity OLED with optimal optical and electricalcharacteristics. We propose here a design procedure that splits the design process into two design stageswhere each stage can be optimized independently. In the first stage we design the OLED withoptimal electrical and optical performance, where the mirrors are specified by their optimal spectral re-flectivity, transmissivity, absorptance, and the phase shift on reflection. In the second stage we synthesizethe top and the bottom multilayer mirrors with a minimal number of layers that satisfy the requiredoptimal spectral dependencies determined in the first part of the design process. As a case studywe present an optimized design for a top-emitting OLED with a simple bilayered cavity consistingof N, N 0-di(naphthalene-1-yl)-N, N 0-diphenylbenzidine (NPB) as the hole transport layer andtris(8-hydroxyquinoline)aluminium (Alq3) as the electron transport layer. Conventional devices withan ITO–LiF=Al electrode pair and a Ag─Ag electrode pair are used as reference devices to benchmarkthe performance of our design. Electrical simulations using the drift-diffusion model and optical simula-tions employing the integrated dipole antenna approach are implemented to test the performance of thedevices. The optimized device shows improved optical and electrical performance when compared withthe reference devices. © 2009 Optical Society of America

OCIS codes: 310.4165, 250.0250.

1. Introduction

Recent progress in the field of electroluminescence(EL) from organic materials suggests that theorganic light emitting diode (OLED), with itsadvantages in size, power consumption, broad colorgamut, high contrast and brightness, and cost, mightseriously challenge conventional display technology[1–3]. Since the discovery of the first efficient organiclight emitting devices by Tang and VanSlyke ofEastman Kodak, intense effort has been put intothe field of research that focuses on improving OLEDperformance [4].To further improve the OLEDs for use in practical

applications, device design and optimization meth-ods that take into account both the electrical and

the optical properties are required. The electricaland optical OLED models have been documented se-parately in numerous studies [5–7]. However, com-prehensive device simulations that combine the twomodels and present optimized OLED device designprocedures have been scarce. In this work, we reporta novel approach for the optical and electrical designand optimization of microcavity OLEDs.

The well-known optical microcavity with metallicmirrors is a simple structure for modifying the angu-lar and spectral distribution of the emitted light,thus leading to improvement in light extractionand reduction of spectral width [8–11]. However,commonly used metallic electrodes such as silver(Ag) typically display relatively poor carrier injec-tion characteristics when compared with pairedindium-tin-oxide (ITO) anodes and lithium-fluoride/aluminum (LiF=Al) cathodes—the most commonly

0003-6935/09/122282-08$15.00/0© 2009 Optical Society of America

2282 APPLIED OPTICS / Vol. 48, No. 12 / 20 April 2009

used anode–cathode pair for devices with N, N 0-di(naphthalene-1-yl)-N, N 0-diphenylbenzidine (NPB)and tris(8-hydroxyquinoline)aluminum (Alq3).Multilayeredmirrors offer the possibility of enhan-

cing the cavity effect while also keeping ideal carrierinjection properties. Several groups have shown thatusing an Ag capping layer on top of a thin LiF=Alcathode can improve the reflectance of the topmirror,and thus the light extraction from the microcavitycan be improved, and at the same time the device dis-plays the desired scheme of having electron injectionfrom LiF=Al [12–14].The conventional OLED design and optimization

methods are either confined to simple structuresor the optimization is performed with respect to onlyone output parameter. To our knowledge, any sys-tematic way of performing multivariable and multi-objective optimization on microcavity OLEDs isalmost nonexistent in the open literature.Most optimization algorithms are currently lim-

ited in terms of finding the optimal solution for afixed number of parameters given certain con-straints. This implies that in the optimization pro-cess the structure of the OLED needs to remainconstant; the variation during the optimization pro-cess is limited to the thickness of the layers withinthe structurally immutable device. If we want to ex-plore the effect of changing the number of layers inthe multilayer mirrors, the optimization needs to beperformed again each time we add or remove a newlayer to or from the mirror. The problem becomeseven less tractable when there is more than onechoice of suitable materials for the new layer. Aset of optimization runs would have to be performedfor every new material.In this work we approach the design and optimiza-

tion process from a different perspective. We proposehere to split the design process into two design stagesthat can be optimized independently. In the firststage we design the OLED with optimal electricaland optical performance, where the mirrors are spe-cified by their optimal spectral reflectivity RðλÞ,transmissivity TðλÞ, absorptance AðλÞ, and the phaseshift on reflection, φðλÞ. In the second stage wesynthesize the top and the bottommultilayer mirrorswith minimal number of layers that satisfy the opti-mal spectral dependencies for R, T, A, and φ, deter-mined in the first part of the design process. Thus ifwe, for example, want to maximize the emitted opti-cal output power and minimize the spectral shiftwith viewing angle, in the first part of the design pro-cedure we will generate the spectrally dependentsets of mirror parameters (R, T, ;A, and φ) that willgive the optimum output characteristics. In the sec-ond part of the design process these parameters formthe input in the mirror design procedure, and the op-timization process is now aimed at designing twomirrors with theminimal number of layers satisfyingthe given spectral target functions.Because of the split between the cavity electrical–

optical design and the mirror design, this new design

procedure is significantly faster, especially for multi-layer mirrors; finding mirror parameters for opti-mum output involves just one optimization run. Theprocedure also offers the advantage of flexibility inhandling the variable number of layers and thechoice of materials, as the optimization process em-ployed automatically searches through the availablematerials database for extra layers and stops wheninsertion of new layers no longer improves perfor-mance. The final case-study design presented in thispaper is a ten-layer top-emitting OLED with optimaloptical and electrical performance for a given set ofmaterials in the design parameters.

We begin by introducing the electrical and opticalsimulation models in Sections 2 and 3. In Section 4we evaluate the conventional device with ITO–LiF=Al and Ag–Ag anode–cathode pairs and use these si-mulation results as the reference to test the perfor-mance of our optimal ten-layer design. Finally, theoptimization procedure, the resulting design, andits performance are discussed in Section 5.

2. Electrical Model

The electrical transport inside the OLED is modeledby the one-dimensional time-independent drift-diffusion model [7,15,16], which solves for a self-consistent solution of electron density n, hole densityp, and potential ψ , implemented in the semiconduc-tor solver Atlas [17]. The basic model includes thecontinuity equations for n (electrons) and p (holes):

ddx

�−μnn

dψdx

þDndndx

�¼ R;

ddx

�μpp

dψdx

þDpdpdx

�¼ R; ð1Þ

where μn and μp are the electron and hole mobilities,Dn and Dp are diffusion constants for electrons andholes, respectively, and R is the recombination rate.The recombination rate is taken to be optical onlyand is modeled by the Langevin recombination coef-ficient γ:

R ¼ γðpn − n2i Þ; γ ¼ 4πeμR

εrε0; ð2Þ

where e is the electron charge, ni is the intrinsic con-centration, and μR is the effective recombination mo-bility, taken to be the larger of the electron and holemobilities in the material; εrε0 is the permittivity ofthe material. Another equation needed is the Poissonequation,

d2ψdx2

¼ −e

εrεo½pðxÞ − nðxÞ þND −NA�; ð3Þ

whereND andNA are the ionized donor and acceptordopant concentrations. The continuity equations andthe Poisson equation are solved to obtain the carrierconcentrations, electric field distributions, and

20 April 2009 / Vol. 48, No. 12 / APPLIED OPTICS 2283

recombination rate. The thickness of the recombina-tion region can be determined from the spatial distri-bution of the recombination rate, which is later usedto estimate the width of the emission region (takinginto account the exciton diffusion). This informationrepresents the input of the optical model.

3. Optical Model

The resonant modes of a microcavity have to satisfythe condition that the phase change during oneround trip is a multiple of 2π. In other words, for nor-mal incidence the following equation holds:

Xi

4πλ diniðλÞ − φtopð0; λÞ − φbotð0; λÞ ¼ 2mπ; ð4Þ

where λ is the emission wavelength, φtopðλÞ, φbotðλÞare the wavelength-dependent phase changes onreflection from top and bottom mirrors, respectively,m is an integer that defines the mode number, andthe summation is performed over all the layers insidethe cavity with thicknesses di and refractive indicesniðλÞ. The phase shift on reflection from the mirrorsis calculated by using the transfer matrix meth-od [18].The intensity of the light output in the form of the

EL spectrum was calculated by using the methodbased on the equivalence between the probabilityof photon emission and the power radiated by a clas-sical dipole antenna [19–21].If we imagine the microcavity with two multilayer

mirrors, in a layer with complex refractive index Ni,the amplitude of the wave vector is given by

ki ¼2πλ Ni; ð5Þ

where λ is the emission wavelength. Opticalconstants of all the materials employed here havebeen modeled by using the Lorentz–Drude model[22,23]. The complex refractive index has been ex-tracted from the variable-angle ellipsometric databy using the simulated annealing algorithm de-scribed earlier [24].When considering wave propagation through thin

film layers, it is convenient to resolve the wave vectorinto two components: the component normal to thedirection of propagation, and the component alongthe direction of propagation (we will use the z axisas the axis of propagation), which is given by

κ ¼ ki sinðαiÞ; ð6Þ

kz;i ¼ ki cosðαiÞ ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffik2i − κ2

q; ð7Þ

respectively, where αi is the angle the wave makeswith the z axis.The total power F emitted by a dipole antenna lo-

cated within the multilayer structure normalized to

the power output of the same dipole in an infinitemedium is given by

F ¼Z

∞

0KðκÞdκ2; ð8Þ

where K is the power density per unit dk2.The power density (K) can be resolved into TM and

TE components, with each component separated intopower densities for dipoles oriented parallel and per-pendicular to the z axis. With this in mind, the powerdensities can be defined as

KTM⊥

¼ 34R

� κ2k2e kz;e

ð1 − aTMþ Þð1 − aTM− Þ

1 − aTM

�; ð9Þ

KTE⊥

¼ 0; ð10Þ

KTMjj ¼ 3

8R

�kz;ek3e

ð1þ aTMþ Þð1þ aTM− Þ

1 − aTM

�; ð11Þ

KTEjj ¼ 3

8R

�1

kekz;e

ð1þ aTEþ Þð1þ aTE− Þ

1 − aTE

�; ð12Þ

where the subscript e denotes the emissive layer(sandwiched between the top and the bottom mir-rors) and a is the reflection coefficient of the mirrorwith respect to the location of the dipole, defined as

aTM=TEþ=− ¼ rTM=TE

e;þ=− expð−2jkz;ezþ=−Þ; ð13Þ

aTM=TE ¼ aTM=TEþ aTM=TE

− : ð14Þ

Here zþ 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 thetop and bottom mirrors calculated by using the mod-ified transfer matrix approach of Katsidis and Siap-kas [25,26].

For a randomly oriented dipole antenna (equalprobability for all orientation directions), the powerdensity is given by

KTM;TE ¼ 13KTM;TE

⊥þ 23KTM;TE

jj ; ð15Þ

where the overall power density will then be theaverage of the TM and TE components.

Clearly, this model can be used in two differentways: treat the complete device as a multilayer struc-ture, or treat the distributed Bragg reflector mirrorsas separate entities specified by their R, T, A, and


φ. Figure 1(a) illustrates the first approach; the op-tical constants and the thickness of each layer areused to calculate the optical output power. However,the same result can be obtained by replacing themul-tilayered top and bottom mirrors by their respectivemirror parameters [Fig 1(b)]. This reduces the com-putational effort and provides a simple and effectivemeans for optimizing microcavity OLEDs withmulti-layer mirrors, since the number of variables has beenreduced to eight (four for the bottom and four for thetop), regardless of the number of layers used in themirrors. Mirrors are treated in the independent op-timization process.By optimizing one or more OLED output optical

characteristics with respect to the mirror para-meters, we obtain the spectral distribution of theparameters for the top and bottom mirrors for opti-mal OLED performance.

4. Results for Reference Devices

The devices shown in Fig. 2 will be used as referencedevices. The optimum thickness of the cavity was de-termined by using cavity equation (4); the thicknessis determined to be approximately 120nm with acavity order of m ¼ 0. The optical field intensity dis-tribution was then calculated for the device in orderto align the antinode of the field with the light emis-sion region of the device. For both devices, the anti-node is located at approximately the middle of thecavity. To align the center of the recombination layerwith the antinode, we used a device with 60nm NPBand 60nm Alq3.The electrical simulation was performed, and the

recombination rates obtained for devices 1 and 2are shown in Fig. 3. It can be seen that, as expected,the device with the ITO anode and LiF=Al cathode

displays a much higher (approximately 4 orders ofmagnitude) recombination rate.

The width of the recombination region obtainedfrom the electrical simulations, which is determinedto be approximately 5nm for both devices, was usedin the optical model to calculate the EL intensity. Therecombination region thickness determines the loca-tion and spread (in terms of width within the cavity)of the spontaneous light emission, which affects theamount of coupling between the spontaneouslyemitted light and the internal optical field. Figure 4shows the EL intensity spectra for emission in thedirection perpendicular to the substrate plane (zaxis). The EL is calculated for a normalized distribu-tion of dipoles along the depth of the recombinationregion. It can be seen from Fig. 4 that by replacingITO and LiF=Al with Ag mirrors, the coupling of thelight within the cavity improves, and thus externallight extraction efficiency (the number of emittedphoton versus the number of photon generated with-in the cavity) is increased by approximately 3 times(in terms of peak value). The spectral half-width isalso reduced from approximately 80nm to less than50nm, thus increasing the color purity.

However, owing to the poor carrier injection fromsilver electrodes (thus the low recombination rate),the total light intensity obtained from the Ag–Ag de-vice is lower than that from the ITO–LiF=Al device.This is the result of the low internal quantum effi-ciency of this device—the number of photons gener-ated inside the cavity per number of carrier pairsinjected is significantly smaller than that observedin device 1. We can conclude that any device designthat tries to take advantage of the microcavity effectmust also achieve recombination rates close to that of

Fig. 1. Equivalence between calculating EL (a) with optical index and thickness (for each layer) of multilayer mirrors and (b) with mirrorparameters R, T, A, φ.


the ITO–LiF=Al device to improve the overall outputlight intensity.

5. Optimization Procedure and Results

The basics of the design procedure are as follows. Inthe first step we maximize the light emission fromthe microcavity OLED by varying the mirror para-meters, namely, mirror reflectance RðλÞ, transmit-tance TðλÞ, absorptance AðλÞ, and phase shift onreflection, φðλÞ. Once the optimal values of thesespectral functions are determined, we proceed to de-termine the multilayer mirror structure (step two)that will provide the required set of RðλÞ, TðλÞ,AðλÞ, and φðλÞ and therefore the maximum light ex-traction efficiency for the device.To demonstrate how our device design procedure

handles multiple design objectives, we include herethe reduction in spectral shift with viewing angle asa secondary design objective. For the reduction ofspectral shift with viewing angle, we note that thespectral shift is approximately given by [27,28]

Δλ ¼Xϑ0

�Xi

4πdi

λ ni½cosϑi − 1� þΔφtop þΔφbot

�;

ð16Þ

where ϑ0 is the viewing angle. It can be seen that byreducing the phase shift on reflection for all angles,we can reduce the spectral shift with viewing angle.

Having established the output light intensity andspectral shift with viewing angle as the objectives ofour optimization, we proceed with the design process.We know that the reflectance for the nonemissionside mirror should be as close to 1 as possible. Thekey now is to establish the optimal characteristics ofthe emission side mirror.

Figure 5 shows the output optical intensity versustop mirror R and T plot at the central design wave-length. It can be seen that the intensity increaseswith increasing R and decreasing T, provided thatthe absorption [1 − ðRþ TÞ] is close to zero. Thus wewant the reflection to be as high as possible withouthaving zero transmission. However, it is difficult toachieve zero absorption in practice, so choosing 90%reflectance and 10% transmittance as the design cri-teria will be technically more feasible. As a result ofthis observation, we will use 10% transmittance asour criteria for transmittance. This conclusion is in

Fig. 2. Device 1 is the simple ITO–LiF=Al bottom-emitting de-vice. Device 2 is the double metallic mirror top-emitting device.

Fig. 3. (Color online) Recombination rates for devices 1 and 2.

Fig. 4. (Color online) Electroluminescence (EL) plot for the refer-ence devices (device 1 is the ITO–LiF=Al reference device, device 2is the Ag–Ag device).

Fig. 5. Extracted light intensity with respect to T and R (whitedenotes maximum or 1.0, black denotes minimum or 0, and eachgray level denotes a 0.1 step).


agreement with the design criteria proposed by Be-nisty et al. [11]. In their paper, Benisty et al. also re-marked that it is technologically more realistic toaim for R of 80%–90% (thus 10%–20% transmit-tance) and minimal absorption. The bottom mirrorused here is 70nm Ag and is fixed.Figure 6 shows the contour plot of the optical out-

put versus the real and imaginary parts of the reflec-tion coefficient of the top mirror, while the calculatedtransmittance is close to 10%. It should be noted thatreflectivity is calculated as the square root of the sumof the square of the real part of the reflection coeffi-cient and the square of the imaginary part of the re-flection coefficient. Reflectance cannot be greaterthan 1; thus the boundary condition is that reflec-tance be smaller than or equal to 1. This explainsthe shape of the contour in Fig. 6. In Fig. 6, it canbe seen that we want the imaginary term to be withinthe region 0–0.2 and the real term to be within theregion 0.8–0.9. We performed numerical optimiza-

tion to determine the exact combination that wouldyield maximum output light intensity, which in turnyields the real and imaginary part of the reflectivityequal to (0.91502, 0.146239). This corresponds toR ¼0:93 and φ ¼ 0:16 rad. Thus we will use R ¼ 0:9,T ¼ 0:1, A ¼ 0, and φ ¼ 0:16 rad (or 9°) as our targetfor the emission side mirror design.

For the mirror design and optimization, we usedthe optical thin film design software OptiLayer [29].OptiLayer uses the matrix formalism for optical mul-tilayer calculations [18,30] and employs an efficientmultiobjective optimization algorithm proposed byFurman and Tikhonravov [30]. The software allowsthe user to specify the “target functions” in the formof RðλÞ, TðλÞ, AðλÞ, and φðλÞ; then the design can berefined by the needle optimization procedure [30].

The materials used for the mirror design proce-dure include aluminum (Al), Ag, ITO, LiF, and NPB.For the nonemission side mirror (bottom), we use a70nm Ag film located between the substrate and theITO anode to boost the bottommirror reflection whilekeeping the ITO as the hole injector. The needle op-timization procedure was used and produced a five-layer top mirror design.

The resulting design can be seen in Fig. 7. An ITOanode and LiF=Al cathode are the ideal electrodes ascan be seen in Fig 2. By keeping the ITO and Allayers thin, we can then use Ag and/or a combinedmetallic and dielectric multilayer to boost the micro-cavity light coupling effect. This can be validated bythe EL comparison between device 1 and device 3shown in Fig 8. Onemight wonder why the optimizeddevice 3 does not perform as well as device 2 (whencomparing Figs. 4 and 8). It should be noted that theoptical simulation does not take the magnitude of therecombination rate into account, thus it shows onlythe optical performance. Because the recombinationrate of device 2 is approximately 6 orders of magni-tude lower than the recombination rate of device 1,device 2 will have a much lower output lightintensity in terms of device luminance when takingboth electrical and optical measures into account.

Fig. 6. Light intensity versus real and imaginary part of reflec-tion coefficient (white denotes maximum or 1.0, black denotesminimum or 0, and each gray level denotes a 0.1 step).

Fig. 7. Resulting structure from the design and optimization pro-cedure (device 3).

Fig. 8. (Color online) Comparison between EL from the optimizeddevice (device 3) and ITO–LiF=Al device (device 1).


However, since device 3 employs the same electrodesas device 1, the optical gain (as seen in Fig. 8) isachieved without compromising the recombinationrate. So the design (device 3) keeps the ideal carrierinjection and recombination properties while boost-ing the maximum output optical intensity by almosttwo times.Another OLED performance measure is the

amount of spectral shift with viewing angle. Suchan unwanted characteristic is prominent in high-Qmicrocavities such as the Ag–Ag device. From Fig 9we can see that the optimized device also displays areduced spectral shift with viewing angle comparedto the Ag–Ag device (device 2). The wavelength shiftfrom the normal direction is a measure of the amountof color change versus the change in viewing angle. Areduction in the wavelength shift versus viewing an-gle means that the CIE (hue) point does not changeas much when we view the device from differentangles.

6. Conclusion

We present a multivariable and multiobjectiveOLED design and optimization procedure that startsfrom a set of materials and in a two-stage processproduces a microcavity OLED with optimal opticaland electrical characteristics. The OLED design re-sulting from the optimization process combines theoptimal electrical performance of the ITO–LiF=Alanode–cathode pair and enhances the output lightcoupling by employing metallic and dielectric multi-layer mirrors. Furthermore, the design shows areduction in spectral shift with viewing angle com-pared with conventional metallic microcavity OLEDswith Ag–Ag mirrors, demonstrating an efficient mul-tiobjective optimization process without increasingthe computation time.

References1. W. Brutting, S. Berleb, and A. G. Muckl, “Device physics of

organic light-emitting diodes based on molecular materials,”Org. Electron. 2, 1–36 (2001).

2. L. S. Hung and C. H. Chen, “Recent progress of molecularorganic electroluminescent materials and devices,” Mater.Sci. Eng. R. 39, 143–222 (2002).

3. J. N. Bardsley, “International OLED technology roadmap,”IEEE J. Sel. Top. Quantum Electron. 10, 3–9 (2004).

4. C. W. Tang and S. A. VanSlyke, “Organic electroluminescentdiodes,” Appl. Phys. Lett. 51, 913–915 (1987).

5. B. Ruhstaller, T. Beierlein, H. Riel, S. Karg, J. C. Scott, andW. Riess, “Simulating electronic and optical processes in mul-tilayer organic light-emitting devices,” IEEE J. Sel. Top.Quantum Electron. 9, 723–731 (2003).

6. C. C. Lee, M. Y. Chang, Y. D. Jong, T. W. Huang, C. S. Chu, andY. Chang, “Numerical simulation of electrical and optical char-acteristics of multilayer organic light-emitting devices,” Jpn.J. Appl. Phys. 43, 7560–7565 (2004).

7. M. A. Webster, J. Auld, S. J. Martin, and A. B. Walker, “Simu-lation of the external quantum efficiency for bilayer organiclight emitting devices,” Proc. SPIE 5214, 300–309 (2004).

8. K. Neyts, P. De Visschere, D. K. Fork, and G. B. Anderson,“Semitransparent metal or distributed Bragg reflector forwide-viewing-angle organic light-emitting-diode microcav-ities,” J. Opt. Soc. Am. B 17, 114–119 (2000).

9. K. A. Neyts, “Simulation of light emission from thin-film mi-crocavities,” J. Opt. Soc. Am. A 15, 962–971 (1998).

10. H. Riel, S. Karg, T. Beierlein, W. Rie, and K. Neyts, “Tuningthe emission characteristics of top-emitting organic light-emitting devices by means of a dielectric capping layer: anexperimental and theoretical study,” J. Appl. Phys. 94,5290–5296 (2003).

11. H. Benisty, H. De Neve, and C. Weisbuch, “Impact of planarmicrocavity effects on light extraction—part I: basic conceptsand analytical trends,” IEEE J. Quantum Electron. 34,1612–1631 (1998).

12. L. S. Hung, C. W. Tang, M. G. Mason, P. Raychaudhuri, andJ. Madathil, “Application of an ultrathin LiF=Al bilayer inorganic surface-emitting diodes,” Appl. Phys. Lett. 78,544–546 (2001).

13. C. W. Chen, P. Y. Hsieh, H. H. Chiang, C. L. Lin, H. M. Wu,and C. C. Wu, “Top-emitting organic light-emitting devicesusing surface-modified Ag anode,” Appl. Phys. Lett. 83,5127–5129 (2003).

14. Y. Q. Li, J. X. Tang, Z. Y. Xie, L. S. Hung, and S. S. Lau, “Anefficient organic light-emitting diode with silver electrodes,”Chem. Phys. Lett. 386, 128–131 (2004).

15. S. J. Martin, G. L. B. Verschoor, M. A. Webster, andA. B. Walker, “The internal electric field distribution in bilayerorganic light emitting diodes,” Org. Electron. 3, 129–141(2002).

16. C. D. J. Blades and A. Walker, “Simulation of organic light-emitting diodes,” Synth. Met. 111, 335–340 (2000).

17. ATLAS User’s Manual—Device Simulation Software (SILVA-CO International, 2005).

18. A. H. Macleod, Thin Film Optical Filters (Institute ofPhysics, 2000).

19. W. Lukosz, “Theory of optical-environment-dependent sponta-neous-emission rates for emitters in thin layers,” Phys. Rev. B22, 3030–3038 (1980).

20. W. Lukosz, “Light emission by multipole sources in thinlayers. I. Radiation patterns of electric and magnetic dipoles,”J. Opt. Soc. Am. 71, 744–754 (1981).

21. W. Lukosz and R. E. Kunz, “Light emission by magnetic andelectric dipoles close to a plane interface. I. Total radiatedpower,” J. Opt. Soc. Am. 67, 1607–1615 (1977).

22. A. D. Rakic, A. B. Djurisic, J. M. Elazar, andM. L. Majewski, “Optical properties of metallic films forvertical-cavity optoelectronic devices,” Appl. Opt. 37,5271–5283 (1998).

Fig. 9. (Color online) Wavelength shift from 0° versus angle (de-vice 2 is the Ag–Ag device and device 3 is the optimized device).


23. J. Chan, A. W. Lu, C. H. Cheung, A. M. C. Ng, A. B. Djurisic,Y. T. Yeow, and A. D. Rakic, “Cavity design and optimizationfor organic microcavity OLEDs,” Proc. SPIE 6038, 603824–1(2006).

24. A. B. Djurisic, A. D. Rakic, and J. M. Elazar, “Modeling theoptical constants of solids using acceptance-probability-controlled simulated annealing with an adaptive movegeneration procedure,” Phys. Rev. E 55, 4797–4803(1997).

25. C. C. Katsidis and D. I. Siapkas, “General transfer-matrixmethod for optical multilayer systems with coherent, partiallycoherent, and incoherent interference,” Appl. Opt. 41,3978–3987 (2002).

26. C. L. Mitsas and D. I. Siapkas, “Generalized matrix methodfor analysis of coherent and incoherent reflectance andtransmittance of multilayer structures with rough surfaces,

interfaces, and finite substrates,” Appl. Opt. 34, 1678–1683(1995).

27. A. B. Djurisic and A. D. Rakic, “Organic microcavitylight-emitting diodes with metal mirrors: dependence of theemission wavelength on the viewing angle,” Appl. Opt. 41,7650–7656 (2002).

28. A. B. Djurisic and A. D. Rakic, “Asymmetric Bragg mirrors forthe reduction of emission wavelength dependence on the view-ing angle in organic microcavity light emitting diodes,” Opt.Commun. 236, 303–311 (2004).

29. DeBell Design, OptiLayer User’s Guide for Windows ’95/’98 &NT/4.0 (DeBell Design, 1997).

30. A. V. Tikhonravov, M. K. Trubetskov, and G. W. DeBell,“Application of the needle optimization technique to thedesign of optical coatings,” Appl. Opt. 35, 5493–5508(1996).


Design of microcavity organic light emitting diodes with optimized electrical and optical...

Documents

Transcript of Design of microcavity organic light emitting diodes with optimized electrical and optical...