HOW To optimize the CIE coordinates of a blue oled by fine tuning the cavity

One goal of emitting material providers for OLEDs in display applications is to synthesize fluorescent, phosphorescent or TADF materials with narrow emission color and a peak wavelength below 450 nm in order to match the requested blue color points of display manufacturers.

In this post we use the optical module of Setfos in order to demonstrate how much the color point of a full OLED can be tuned for a given emitter spectrum. Naturally, this comes with the drawback that we cut some of the PL emission spectrum by tuning the OLED cavity. Therefore, we also compare the efficiency of the color-optimized device with the one that is tuned for highest outcoupling/radiance and analyse the consequent trade-off between color purity and efficiency. Note that the same analysis and optimization can be performed for other types of emitters such as quantum dots or perovskites. Despite the more narrow emission spectra of these emitters, the optics of such QLEDs and PeLEDs is equal to the well established OLEDs and the cavity effects are thus of outmost importance.

Introduction

An important key figure for a display is the color range that it covers. It is typically specified by a percentage coverage of specific standard color spaces. There are various standard color spaces (or gamuts) depending on the specific application of the display (see Figure 1). Two important ones for TV screens are the DCI-P3 and the Rec.2020 (or BT.2020) standard. While the first one is already fully covered in some prototype displays (https://www.oled-info.com/here-are-joleds-new-oled-display-prototypes) and likely to come for commercially available TVs soon, the latter is defined by three color points that are equivalent to monochromatic light sources (on the CIE spectral locus) and is thus really challenging (if not impossible) to achieve. The larger the coverage of the color space, the purer are the emitted colors and the better the experience for the viewer. The first prerequisite for achieving a wide color range with the display is clearly a narrow emitter spectrum. However, thanks to thin-film interference effects in OLEDs, QLEDs, or PeLEDs, peak wavelengths can be shifted and the full-width-half maximum of the emission spectrum can be shaped considerably by tuning various layer thicknesses. In this blog post, we start with two organic deep blue emitters presented recently[1] and employ them in standard bottom- and top-emitting OLED structures. By tuning selected layer thicknesses we demonstrate the effect of the cavity on the blue emission color, analyze the trade-off between optimal color and best efficiency, and try to achieve the blue color points of the DCI-P3 and the Rec.2020 color spaces by optimizing the stack.

Here, we are presenting how to perform optical and electrical simulations on blue OLEDs by using the software Setfos.

Fig. 1. Different color spaces used for monitors (https://www.benq.eu/en-is/knowledge-center/knowledge/color-gamut-monitor.html)

Setting up the Simulation

In a recent work by the Zysman-Colman group [1], two deep blue thermally activated delayed fluorescent (TADF) emitters have been synthesized, characterized and employed in OLED devices. Their solid-state PL spectrum (Figure 2), the PLQY and the prompt decay lifetime (Table 1) are used in the following as simulation input in Setfos.

Table 1: Characteristics of the blue emitter determined from solid state PL: prompt decay lifetimes, PLQY, peak wavelength, FWHM and color points

For the first analysis, we use a standard bottom-emitting OLED that is typically used in research labs to characterize new emitters. The prototypical structure is shown in Figure 2. Besides the layer structure and the refractive index of each layer we also need to define the emitter properties of the EML in Setfos. For simplicity, we assume a delta-emitter in the center of the EML and use an isotropic emitter orientation. Further input parameters are the quantum efficiency and the lifetime for the emitting material which are specified in Table 1.

Optical simulation results

From an optical point of view, all layers (with their respective thickness and the refractive indices) will influence the cavity. Therefore, it is in principle possible to adapt all those parameters to optimize the stack for improved outcoupling, color point, reduced angular color shift, etc.

Here we restrict the free parameters to the ETL and HTL thickness. In order to illustrate the effect of these two parameters, we sweep them in Setfos and check the emitted color and the efficiency of the used OLED. Note that also for an optimization of the cavity, the knowledge about the emitter position (“emission zone”) is essential. To find further information about the extraction of this essential parameter, you can use the Phelos instrument page as a resource.

Figure 3a shows the emitted color for varied ETL and HTL layer thickness. It is apparent that especially the ETL layer thickness is influencing the emitted color greatly. From purple, over blue, and sky-blue to greenish, all colors can be obtained. To illustrate these various colors in the CIE color diagram, we also swept only the ETL thickness for a fixed HTL thickness as indicated by the yellow arrow in Figure 3a. The resulting color points are shown in Figure 3b. The dominant effect of the ETL thickness on the emitted color is also present for the radiance and can be reasoned easily. The transport layer thicknesses are directly influencing the distance between the emitter and the electrodes. These distances are determining the - ideally constructive - interference criteria of the OLED. As the ITO electrode, in contrast to aluminum, is only partially reflecting the emitted light, the distance from the emitter to the ITO electrode (mainly determined by the HTL thickness) is influencing the cavity to a much lower extent compared to the emitter-Al distance (mainly determined by the ETL thickness). The forward emitted radiance is shown in Figure 3c and shows clear maxima for distinct ETL thicknesses, which are separated by about 120 nm (~ peak wavelength / 2*mean refractive index). Alternatively to the radiance in forward direction, one could also analyze the outcoupling efficiency which shows the same trend with layer thicknesses. But as we are focusing our attention on blue OLEDs for display applications, this integrated quantity over the upper hemisphere is less important than the forward (0°) emission intensity. The OLED efficiency is here quantified by the so-called blue index which is calculated from the ratio of the current efficiency (cd/A) divided by the CIEy coordinate. This quantity is especially useful in the display research for blue OLEDs, as it aims to partially compensate for the (deep-)blue OLEDs penalty that is present in the luminance due to the lower eye sensitivity in this wavelength range. For both radiance and blue index a distinct maximum of 11.8 W/(m^2 sr) and 117.52, respectively, can be seen for ETL/HTL thicknesses of ~37.5/70.5 nm. This optimum is in the following used as a reference when we focus on the potentially achievable color points rather than on efficiency.

Figure 3: Sweep of ETL and HTL thickness and resulting emitted color (a), radiance (c) and blue index (d). The color diagram in (b) shows the emitted color along the yellow arrow in (a) which corresponds to an ETL thickness variation only.

Before we move on to the optimization of the color point, we look at some other key quantities calculated in Setfos which are related to the emitted color. To do so, we fix again the HTL thickness to 70 nm and sweep only the ETL thickness as indicated by the yellow arrow in Figure 3a. Figure 4a shows again the radiance, blue index and the emitted color for the specified sweep. In addition, we now also show the CIE coordinates and the full-width-half-maximum (FWHM) of the OLED. The FWHM value can be tuned considerably from ~90 to < 40 nm by changing the ETL thickness. Especially compared to the FWHM value of the pure emitter film of 74.2 nm (see orange dashed line in Figure 4a) this indicates the capabilities of using the OLED cavity. Especially for intrinsically broad emitter materials such as phosphorescent and TADF emitters, this possibility is of outmost importance. The FWHM is typically also a good indication of the color purity of the OLED. However, care has to be taken with the interpretation of this quantity. For this particular OLED the FWHM suggests a desirable color purity above an ETL thickness of ~95 nm (see second dashed line in Figure 4a). However, the CIE color coordinates are not at all in an acceptable range at 100 nm. The reason is shown in the simulated emission spectra in Figure 4b. Besides the main peak, a second peak is appearing due to multiple resonance criteria in the OLED. This peak is considerably influencing the color point of the OLED with an ETL thickness of 105 nm. Only if the second peak is negligible again (~125 nm) the color point reaches its closest proximity to the Planck locus and the OLED emits its “most pure color”.

Figure 4. a) radiance, blue index, color coordinates and FWHM variation as a function of the ETL thickness. The yellow dashed line represents the emitter PL FWHM. b) Normalized emission spectra for selected ETL thicknesses.

To summarize these results shortly, we can optimize the ETL (and HTL) thickness to get the brightest OLED (ETL: 37.5 nm, radiance: 11.8 W/(m^2 sr), CIE x/y: 0.152/0.1) by losing color purity, or we can tune the ETL to achieve the lowest CIE y value (ETL: 124.4 nm, radiance: 4.55 W/(m^2 sr), CIE x/y: 0.16/0.029) with a considerable brightness drawback. In reality, we need to balance this trade-off, we want to tune the OLED for a specific color point while maintaining the highest possible radiance! This is the topic of the next section.
All the illustrative simulations above have been performed with the pDTCz-3DPyS emitter. The same analysis could be performed with the second emitter shown in Figure 1 or with any other emitter spectrum, not restricted to organic molecules.

Optimization for dedicated color points

In this section, we show how the optimization function in Setfos can be used to find the best structure to minimize the trade-off between color point and efficiency. To illustrate the problem we show again the radiance and color coordinates for the ETL and HTL thickness variation (see Figure 5). For the first standard color space – DCI-P3 – the coordinates of the blue color point are given by (0.15/0.06). Therefore we set these values as upper limits in the CIE x and y plot. Those plots are now used as masks for the radiance plot. The task - which can either be performed manually or by using Setfos - is now to find the maximum radiance within the non-shaded area.

Figure 5: Radiance and color coordinate plots with adapted limits CIE x/y < 0.15/0.06. These limits act as a mask for the radiance plot. The task is now to find the maximum radiance within the possible (not shaded) regions. Here the maximum is achieved for about 150 and 75 nm thick ETL and HTL, respectively.

The results of this optimization are shown within Table 2 and compared to a simple color point optimization for both emitters. For both emitters, the DCI-P3 blue color point can be easily reached. Due to the red-shifted and broader PL spectrum of the pDTCz-2DPyS emitter, the loss in radiance is significantly higher. Depending on the tolerance for the color points you can still somewhat optimize the radiance. In Setfos this optimization task is achieved by optimizing color coordinates and radiance at once but with a significantly higher weight for the color point. If tolerance of < 5% is acceptable to the color point, 38 and 87% of the maximum radiance with the BEOLED structure can still be achieved for pDTCz-2DPyS and pDTCz-3DPyS, respectively.

Meeting the first blue color point standard (DCI-P3) seems not to be a problem with both emitters and a simple bottom-emitting OLED structure. In the last part of this blog, we want to analyze whether we can also get closer to the blue color coordinate of the next generation color space defined in Rec.2020. As an initial step, we simply adapted the optimization target to the new color coordinates and rerun the Setfos optimization. However, with the given emitter spectrum and bottom-emitting OLED structure, the required target cannot be met satisfyingly (see Table 2).

Table 2: Summary of the color optimization runs using the BEOLED structure for both emitters. When the optimization is only considering the color targets (DCI-P3), the radiance is 10 – 20% lower. The blue color point of Rec.2020 cannot be reached with the given emitter in the given bottom-emitting OLED

A potential way of still getting closer to the stricter Rec.2020 color point would be to use a stronger cavity. This is usually achieved by employing so-called top-emitting OLED structures (TEOLEDs) which are standard for display applications. Therefore we check the potential of a prototypical TEOLED structure in the following (see Figure 6). For simplicity, we only use the ETL (BmPyPB), HTL (TAPC), and capping layer (TAPC) thicknesses as optimization parameters. Clearly, reaching exactly the blue color point of the Rec.2020 standard is not realistic because the blue color coordinate is equal to a monochromatic light source of 467 nm. Nevertheless, thanks to the TEOLED structure, we can get close to the target value with both emitters (see Table 3).

Figure 6: Used Top-emitting OLED structure used in this simulation

Interestingly, both emitters can be tuned equally well by the cavity to the target color point of Rec.2020, and also the loss in radiance compared to the maximum radiance is very similar for both. This is interesting as most often the peak wavelength of emitters is used as an important quantity. This example demonstrates that the full spectrum needs to be considered in the optimization which is readily done in Setfos. The FWHM values of both emitters are close to each other. We speculate that this similarity explains the similar performance for the Rec.2020 target. In turn, this suggests that matching the Rec.2020 even better would require a more narrow PL emitter spectrum. Such narrow emitter spectra can either be achieved by a fluorescent dopant in an organic matrix (hyperfluorescent™ OLEDs), by quantum dot materials (QLEDs), or by perovskite emitters (PeLEDs).

Table 3: Summary of the color optimization runs using the TEOLED structure for both emitters. The emitted color can be tuned close to the blue color point of Rec.2020 but a considerable loss in radiance is seen.

Conclusions

The OLED cavity cannot only be used to optimize the radiance of an OLED but also to tune the emitted color. Here we used this possibility to match the blue color points of the two standard color spaces DCI-P3 and Rec.2020 for two given blue TADF-based organic emitters. The first standard DCI-P3 can be reached fairly easily with both emitters. For the emitter with the lower peak wavelength, we can even resemble 87% of the maximum radiance in a bottom-emitting OLED configuration. The monochromatic Rec.2020 color point can clearly not be reached in the traditional bottom-emitting OLED. Using a top-emitting OLED provides an acceptable alternative to get closer to the requested color point. However, the loss in radiance is significant. Future display applications, therefore, concentrate on the use of narrower emitter spectra, be it as fluorescent dopant, quantum dot, or perovskite material. But also for these new types of emitters the stack engineering, similar to this blog post, stays important.

References:

• [1] P. Rajamalli, D. Chen, W. Li, I.D.W. Samuel, D.B. Cordes, A.M.Z. Slawin, and E. Zysman-Colman, J. Mater. Chem. C 7, 6664 (2019).

Back to the top