The external quantum efficiency (EQE) described the number of outcoupled photons per injected charges and is defined by the 4 factors charge balance, singlet-triplet generation efficiency, radiative quantum efficiency and light outcoupling efficiency:


In an OLED stack, only part of the emitted light can escape the device, this problem is caused by a high refractive index in organic layers. Most of the light is lost to three loss channels ( for a top-emitting OLED ):

  • guided modes ( total internal reflection )
  • absorption losses ( absorption in layers and at electrodes interface )
  • surface-plasmon-polariton ( evanescent coupling mode )

In this study, we optimize the outcoupling efficiency by changing layers material, emitter properties and electrode material.

  1. Emitter characterization with PHELOS measurement and SETFOS fitting
  2. Parameters of the simulation
  3. Impact of layers refractive index
  4. Impact of dipole distribution
  5. Impact of dipole orientation
  6. Impact of electrode material and thickness


1. Emitter characterization with PHELOS measurement and SETFOS fitting

To optimize the light extraction efficiency, we have to know some properties about the OLED, like layer thicknesses, emitter position, distribution, orientation and emission spectrum. By using two products Phelos and Setfos, we can get this information.

Tutorial for Phelos measurement coupled to Setfos fitting

Here is our reference stack in figure 1  which is a  top-emitting OLED, we will change different parameters in Setfos to see their impact on the outcoupling efficiency.

reference stack.png
Figure 1: Reference stack, top emitting OLED

Back to top

2. Parameters of the simulation

For the case of a bottom-emitting OLED ( figure 2 ), to escape the device light has to go through the thin transparent ITO electrode and the glass substrate. As the organic layer has a high refractive index about 1.8, much higher than the glass (about 1.5) and the air (n=1), there are two total internal reflections in a bottom-emitting device. Light can escape the device only if it is emitted within the escape cone which is defined by materials refractive index, the incidence angle of the emission is defined by the dipole orientation.

bottom electrode.png
Figure 2: Bottom emitting OLED

For the case of a top-emitting OLED ( figure 1 ), light can be reflected by the top electrode and by the bottom electrode, the constructive interference condition between emitted light and reflected light depends on the wavelength, the incidence angle of the emission which depends on the dipole orientation, the ETL, HTL thicknesses, the dipole position and the layers refractive index.

For a top-emitting OLED usually a capping layer (CPL) is used to increase the outcoupling efficiency by tuning the CPL thickness. The CPL thickness also influences the emitted color at different angles, as shown in figure 3. 

CPL impact.png
Figure 3: Impact of the CPL layer

For each case in this study, we optimize the outcoupling efficiency firstly with ETL and HTL thicknesses and then with CPL thickness.

We use mode analysis as emission mode in Setfos to analyse in detail the loss channel contributions.

Back to top

3. Impact of layers refractive index

According to a publication by Salehi et al. ("Manipulating refractive index in organic light emitting diodes"), layers refractive index can be reduced by pore inclusion using oblique angle deposition leading to increased light extraction efficiency in OLED devices. For the simulation with Setfos we can use different ETL and HTL materials with different refractive indices to investigate the impact on the outcoupling efficiency.

To do the optimization of extraction efficiency, we use the emission module and the sweep mode. By defining the sweep range, we will get result for each case in the range. Figure 4 shows a plot of the outcoupling efficiency I_OC versus HTL and ETL thicknesses, we can see four maxima. We focus on the first and the second maximum in the following.

Figure 4: Outcoupling efficiency versus ETL and HTL thicknesses

In the next step we optimize the CPL thickness to further improve the first and the second maximum.

Figure 5 depicts simulation results for different cases from the reference stack to the stack where ETL and HTL have lower refractive index material (about 1.65), the outcoupling efficiency I_OC is equal to 26% for the first maximum in the reference stack.

impact of ETL and HTL refractive index.png
n change.png
Figure 5: Mode contribution for ETL and HTL change by material with n~1.65
  • By decreasing the refractive index of both ETL and HTL materials from 1.8 to 1.65, we increase the outcoupling efficiency I_OC by 31%  (compared to the reference stack) in the first maximum and by 25% in the second maximum.
  • By decreasing the refractive index, we decrease considerably the guided mode plus evanescent coupling mode I_GM+EC by 21% for the first maximum and by 30% for the second maximum.
  • The absorption losses I_AL increase for the second maximum, this is why I_OC does not increase a lot whereas I_GM+EC decreases considerably for the second maximum.

According to the publication by Salehi et al., "It has been shown that the refractive index of Alq3 can be tuned from 1.75 to 1.2 using OAD (oblique angle deposition)". We can therefore further increase the outcoupling efficiency (up to 150% compared to the reference case ) by using even lower ETL and HTL refractive index ( n=1.2 ). The question is whether such low-index material can be produced.

lower HTL and ETL n.png
Figure 6: Mode contribution with lower ETL and HTL refractive index

In the next step we investigate the influence of the refractive index of the EML. But it is not the same story because the outcoupling efficiency is higher for high EML index, as seen in figure 7.

lower EML n.png
Figure 7: Mode contribution for lower EML refractive index (n_ETL=n_HTL~1.65)

Back to top

4. Impact of dipole distribution

Constructive interference of reflection at the electrode interface depends on the wavelength, the incidence angle, layers refractive index, layers thicknesses but also on the dipole distribution and position. We consider the low-n (n=1.65) stack with isotropic emission and with optimized layer thicknesses for dipoles with Dirac distribution at position 0, i.e. at the ETL interface. We now change the dipole position from 0 to 1 by steps of 0.1, it changes the outcoupled flux by 22%. These changes are however only because the outcoupling efficiency had been optimized for dipole position 0.

impact of dipole position.png
Figure 8: Mode contribution versus dipole position for layer thicknesses optimized for dipole at position 0

If we optimize layer thicknesses for each dipole position, we find that the outcoupling efficiency does not change ( figure 9 ).

dipole position on the first max.png
Figure 9: Mode contribution versus dipole position with optimized layers thickness (the first maximum)

Therefore the dipole distribution and position must be known to optimize layer thicknesses, we can get this information with Phelos measurement coupling to Setfos fitting.

Back to top

5. Impact of dipole orientation

Emitters can be regarded as dipoles which emit light into a certain direction. In the following we will study the impact of dipole orientation on the outcoupling efficiency. Figure 10 shows two cases of dipoles in an EML layer, a perpendicular dipole which emits light inside the plane of the emitting layer and a parallel dipole which emits light mostly into the escape cone, the parallel dipole is therefore more efficient.


two kinds of dipole.png
Figure 10: Two extreme cases of dipole, the perpendicular one and the parallel one

In Setfos, we quantify dipole orientation by a number between 0 and 1. Hereby 0 means all dipoles are parallel and 1 means all dipoles are perpendicular. We consider the low-n stack with dipole position 0.5 (in the center of the EML) and change the dipole orientation from 0 to 1 by steps of 0.1. For each case we optimize layer thicknesses for the first maximum. The result ( figure 11 ) shows that parallel dipoles are most efficient with a small contribution of guided mode plus evanescent mode because light is almost completely emitted in the escape cone. Between the parallel dipole and the perpendicular one, the optimized layer thicknesses change considerably from 50 nm to 140 nm.

impact of dipole orientation.png
Figure 11: Mode contribution versus dipole orientation

Emission layers with highly parallel dipole orientation will therefore lead to a good outcoupling efficiency. ( We can extract the dipole orientation with Setfos fitting from Phelos measurement data. )

Back to top

6. Impact of electrode material and thickness

We will now consider the low n stack with Dirac distribution at position 0.5, and this time with dipole orientation equal to 0.18 (the result from the fitting). We change top and bottom electrode material between silver (Ag) and aluminum (Al) to study the impact of the electrodes material. Figure 12 shows the difference in absorption and reflection between the two materials. We notice that silver reflects more and absorbs less than aluminum for wavelengths over 420 nm.

difference between Ag and Al.png
Figure 12: Reflectance (solid lines) and absorbance (dashed lines) of Ag and Al versus wavelength

In figure 13 we plot the comparison of four cases on the first maximum and on the second maximum contribution of each mode. It shows that silver as top electrode induces less absorption losses, and the best case is the case with silver as top electrode and aluminum as bottom electrode.

electrode material.png
Figure 13: Mode contribution for four combinations of electrodes material

We can also see the impact of the top electrode thickness. In figure 14 we plot the comparison between cases with top electrode thicknesses of 20, 10 and 5 nm. There is no big impact on the outcoupling efficiency, so we have to consider electrical effects first and not use a too thin electrode. There is no impact of the bottom electrode thickness because it is a very thick layer. Optically, the bottom electrode is like a mirror which reflects light .

electrode thickness.png
Figure 14: Mode contribution for different top electrode thicknesses

In figure 15 we show a summary of the parameter changes employed in our study to increase the outcoupling efficiency. In the reference stack we already use Ag as top electrode. We can demonstrate an increase of the EQE from 26 % to 51 % by optical optimization! We can increase even more if we get lower ETL and HTL index material.

Figure 15: Mode contribution from the reference stack to low n stack ( n~1.65 ) with parallel dipole.

Back to top

To conclude:

  • We used Phelos measurement and Setfos fitting to determine the emitter orientation, emitter position and emission spectrum.
  • We used Setfos to optimize the EQE for HTL, ETL and electrodes material, the dipole orientation and position should be known from Setfos fitting to optimize layer thicknesses.
  • In order to obtain the optimum EQE in OLEDs, we have to use low-index ETL and HTL, parallel oriented emitters and silver as top electrode.

Back to top