Photolithography for High Resolution OLED Display

Modern society has grown accustomed to an overflow of visual information, with displays in the center of most user interfaces. The pace of introducing new technologies and of reducing the cost of manufacturing has been impressive and does not seem to slow down. The most prominent examples are OLED displays (based on organic light-emitting diodes), evolving from curiosity only some years ago to a technology that is dominating the market position today. 2017 has seen a major increase in both shipments (more than 400 million units) and revenue (around $25 billion) for AMOLED display panels (according to UBI Research and DSCC).

From the very beginning of OLED history, it was crucial to find a way to maintain efficient emission in stacks composed of very fragile materials. As most of the materials used in an OLED structure are highly sensitive to a lot of elements (e.g., air, moisture, solvents, temperature, radiation), protecting the device has always been crucial, both during fabrication and during operation. This has evolved into several research tracks. Firstly, great effort by material companies to synthesize new molecules and polymers resulted in many OLED families, both for thermal evaporation and solution processing. Secondly, equipment advances made it possible to uniformly deposit stacks on large substrates with industrial takt time. Thirdly, different encapsulations were developed to protect the OLED stack during usage to ensure enough lifetime for consumer applications. All of the above required years of research and significant investments, which makes it challenging to introduce new OLED manufacturing techniques and change the existing process flows.

At the same time, current manufacturing methods have their limitations. Two main approaches are color-by-white (WOLED) and side-by-side red-green-blue (RGB OLED), differing by the way that the colors are realized in subpixels (FIGURE 1). In WOLED, the light source is a continuous layer of a broadband (white) OLED emitter and the three basic colors are selected by passing the light through color filters (CF). The advantage is that the pixel density is limited only by the backplane resolution and the CF resolution, which is why this is the main concept used for OLED microdisplays with CMOS circuitry. The disadvantage is that a significant portion of the light is lost due to CF absorption, which impacts the display power efficiency. In RGB OLED, each subpixel is a different material stack, so each subpixel is a separate light emitter. This is typically realized by depositing each stack by thermal evaporation through a fine metal mask (FMM) and is used for most smartphone OLED displays. The advantage is that each color is optimized, so the display efficiency is much higher. At the same time, it is difficult to scale the FMM technique both in substrate size (masks tend to bend under their own weight, so the mother glass has to be cut for OLED deposition) and in resolution (standard masks are not suitable for resolutions above several hundred PPI and the cross-fading area limits the aperture ratio).

An alternative way to realize side-by-side RGB pixels is to use photolithography techniques known very well from the semiconductor industry (and used in displays for the TFT backplane fabrication). In such a case, after depositing a blanket OLED stack, photoresists could be used to transfer the pattern and remove the unnecessary material by etching (FIGURE 2). The challenge here is, again, susceptibility of OLED materials to solvents – using standard (semiconductor) photoresist chemistry results in dissolution/removal of the stack. Still, the gains are definitely worth the extra effort, as litho can provide both very high pixel density (submicron pixel pitch) and, at the same time, a very high aperture ratio (emitting area maximized thanks to minimizing pixel spacing). Over the years, some new approaches for photolithography have been proposed. One way, followed by Orthogonal Inc, is to use fluorinated materials which should not have any chemical interaction with the organic stacks (thus, orthogonal to OLED). The other approach, followed by imec together with Fujifilm, is to pattern organic stacks using a non-fluorinated, chemically amplified photoresist system.

For imec, an R&D hub with long traditions of developing new photolithography nodes, organic photolithography is a way to address the challenges of next-generation high-resolution displays. In virtual and augmented reality (VR/AR) applications, the display is very near to the eye of the user. This results in very aggressive requirements in terms of pixel density in order to avoid annoying “pixilation.” The same goes for the required minimum pixel spacing, to avoid the “screen door effect”. With photolithography, these two challenges can be addressed simultaneously. The OSR photoresist system from Fujifilm can deliver lines and spaces with a 1 μm pitch, which fits in the roadmap towards several thousand PPI resolutions for the OLED front plane. We have realized a dot pattern transfer to the OLED emission layer with a 3 μm pitch, which corresponds to 8400 PPI resolution in a monochrome array. After stripping off the photoresist, the EML remains on the substrate, as verified by photoluminescence (FIGURE 3).

On the device level, we have fabricated OLED arrays with 10 μm pixel pitch (FIGURE 4), corresponding to 2500 PPI. In this case, an important parameter is the alignment accuracy, which defines how much of the total display area can be used for emission. Another limitation is the resolution of the PDL (pixel definition layer), a dielectric layer separating the OLED stack from the bottom contact level. The resolution of this layer limits the maximum opening that can be achieved, which translates to the aperture ratio of the pixel – or the percentage of the area that is used for OLED emission. In this example, the “photoluminescence aperture ratio”, or the relation of the OLED island to the pixel area is around 50%, which is enabled by small spacing (<3 μm). However, the “electroluminescence aperture ratio”, of the relation of the area emitting light, is 25% because of the PDL area and the necessary overlap of the OLED island. Assuming minimum line spacing of 1 μm, one can envision a PL ratio of 81% (9 x 9 μm) and EL ratio of 64% (8 x 8 μm) for a subpixel of 10 x 10 μm. With such scaling, the usable area of the array can be enlarged, which results in a longer device lifetime (since we can reduce the driving current density) and in reduction or elimination of the screen-door effects.

Obviously, interrupting the optimum deposition process in ultra-high vacuum and exposing the OLED stack to photolithography materials has an impact on the device performance. Just breaking the vacuum results in a hit on lifetime performance. Additionally, our initial process flow includes exposure of the stack to the ambient atmosphere (air and humidity), as we have been using standard cleanroom equipment. In the beginning, such a “worst-case scenario” resulted in proof-of-concept of emitting OLEDs after patterning, but, unsurprisingly, with a device lifetime of only a few minutes. In the course of the development, we have introduced improvements on three fronts. Firstly, there have been continuous upgrades of the photoresist system to make it more compatible with the organic stack. Secondly, the process flow has been optimized to reduce the impact of process parameters on device performance. Thirdly, the OLED stacks have been tuned for robustness, for example by introducing additional protection layers for the most critical interfaces. All these actions resulted in device lifetimes of several hundred hours at 1000 nit luminance. As the lifetime is the major concern when it comes to the readiness of this technology, this is an ongoing effort to bring all the parameters to a level acceptable by the industry.

In parallel to performance improvement, we have been developing a route for patterning multicolor arrays with photolithography. The main challenge, in this case, is to protect the previous “color” (OLED stack) while patterning the next one. Once this condition is satisfied, side-by-side arrays with several stacks can be realized – and, this is not limited to light emitters. Next to red-green-blue OLEDs, for example, an organic photodetector subpixel could be fabricated to add functionality to the display. In terms of manufacturing, each “color” of the front plane would be fabricated in a similar way as it is done for each layer of the backplane.

In our recent work, we fabricated a 2-color passive OLED display and this prototype was demonstrated at the Touch Taiwan 2017 exhibition (FIGURE 5). The 1400 x 1400 pixel array has a subpixel pitch of 10 μm, resulting in a resolution of 1250 PPI. The stacks are phosphorescent red and green small-molecule OLEDs, deposited by thermal evaporation. The display is designed for top emission and uses glass encapsulation. Thanks to the separate driving of two groups of subpixels, the two colors can be displayed independently. The prototype has been in operation for tens of hours with all pixels turned on, with no visible degradation. This indicates that the process flow for multicolor patterning proves basic functionality and already ensures stability for reasonable working time. A similar front plane can be integrated with a TFT or CMOS backplane, enabling then video mode of operation, with individual driving of each subpixel. In a separate demonstration, we have also verified that the fabrication process is compatible with an FPD backplane process using IGZO TFT and flexible substrate.

Taking everything into account, the photolithography of organic semiconductors is an emerging technology that can enable high-resolution OLED displays. Many technology milestones have been already cleared – we know that we can achieve patterns of few microns, realize side-by-side multicolor pixels, integrate the pixelated front plane on different backplanes, and get encouraging efficiency and lifetime performance. Currently, optimization of OLED performance after patterning is still the top priority. At the same time, we are addressing the complete integration flow and manufacturability aspects. To have this technology fully incorporated in a fab process flow, material and equipment developments are required. Still, the prospect of ultra-high resolution with simultaneous high aperture ratio in a process flow based on standard semiconductor techniques remains very attractive and justifies going the extra mile to tackle the pending engineering challenges. Let us know. Contact us.