MicroLEDs have been on the emerging technology radar for quite some time, with anticipated power efficiency fueling hype around their potential for next-generation displays. Over the past five years, it has become clear what applications microLED can and cannot enable in the near term – and what manufacturing challenges remain to be resolved for mass-market consumer applications to become financially viable. This blog summarizes the key issues for microLED technology and looks at maturity for various applications, as well as the limitations of different approaches that display makers pursue.
Key issues for microLED tech
Those considering microLED opportunities should consider the following:
- The power efficiency of microLED decreases with size. Shrinking LED size and interchip space (pitch) offers higher-resolution applications or lower power consumption for a desired level of brightness. However, LED chips lose 50% power efficiency going from 1,000 μm to 10 μm due to manufacturing blemishes/defects that become increasingly detrimental with smaller LED-substrate interface areas. The issue stems from the structural differences between emissive multilayer materials and substrates, which during the deposition steps results in defects. The smaller the chip area, the greater defect impact. There are potential workarounds, such as using buffer layers to smooth the transition from substrates (silicon, sapphire, etc.) to indium- and gallium-based active materials, or building nanowire-shaped LEDs with a smaller base, but more work is needed to show that these are viable at scale.
- Creating a fully colored microLED display triples the efforts compared to incumbent displays. LEDs are fabricated with a specific emission color. Achieving full-color microLED displays requires individual red, green, and blue (RGB) monolithic LED chips or downconverting light from one common blue LED source to green and red with quantum dots. For the first option, monolithic fabrication conditions differ between the colors, with green requiring higher doping of indium compared to blue indium gallium nitride (InGaN) LEDs and red emission coming from a different composite like aluminum gallium indium phosphide (AlGaInP), which is incompatible with the blue and green LED fabrication process. This approach presents challenges, with variations in indium concentration and deposition conditions affecting epitaxy quality (increasing defects), differing driving voltages/parameters for different colors complicating driver circuitry, and increased surface defects with lower power efficiency as LED chiplet size decreases. The second option, color downconversion using quantum dots (QDs), is promising for larger LEDs, which are easier to fabricate and handle, but QDs are harder to apply to smaller LEDs, where thicker layers are required to absorb the same amount of light. Additional patterning structures are necessary to avoid color/pixel cross-talks (similar to camera CMOS design). There are several ways to pattern QDs, depending on the accuracy needs, with photolithography allowing for several-micron linewidths and thicknesses, electron-beam lithography achieving several-nanometer linewidths and thicknesses of tens of nanometers, inkjet printing providing a 250 nm line and several-micron thicknesses, and dip-pen lithography leading to tens-of-nanometer widths and monolayer thickness.
- MicroLED transfer and placement is still a bottleneck for mass-market adoption. Fabricating densely packed microLEDs on a large substrate is financially effective, but many display applications do not need very high pixel resolutions like 30,000 ppi on the wafer. Taking advantage of wafer-based semiconductor manufacturing and associated capex efficiency means that the microLEDs must be sliced off and distributed across a new substrate for final display applications. Several approaches vary in the combination of yield and throughput. Elastomer transfer/stamping suits small areas and can move 10,000 to 25,000 units per hour. Electrostatic and electromagnetic transfer suits large areas but suffers from low yields due to voltage damage. Laser-assisted transfer has yields of around 90% and can place 100 million units per hour but suffers from chip damage, while fluid self-assembly has low cost but also low yield, placing some 50 million units per hour. Given the variations here, each of the approaches will likely be limited to specific display segments.
Applications for microLED technologies will shift with tech variations
MicroLED technologies across display segments have differing maturity levels. Applications that require large form factors and sparsely placed emissive pixels suffer the most from a low yield of transfer processes and will see adoption only after other display segments become widely available to consumers. On the other hand, augmented reality (AR) and wearable displays can use monolithic and small-size transfer processes and will enter the mass market in the near future, especially as these products are the least sensitive to handling defects. Each display application group differs in resolution, display size, and consequently chip dimensions.
- AR/VR uses around 1-inch panels with 1 µm to 5 µm chips and 500 PPI to 2,000 PPI resolution.
- Wearables can use 1-inch to 1.5-inch panels, with 10 µm to 30 µm LEDs at 200 PPT to 300 PPI resolution.
- Mobile displays need 4-inch to 7-inch panels and 30 µm to 50 µm chips with 300 PPT to 800 PPI resolution.
- Both automotive and TV panels can be in the range of 6 inches to 100 inches and can work with 50 µm to 100 µm LEDs at 50 PPT to 250 PPI – though cars may need a slightly higher resolution, as the viewer is closer to the displays varying in sizes.
While microLED technologies continue to attract investments, both corporate and startup efforts are consolidating around specific display applications, indicating approaching technical maturity. The display industry overcame the hype to realize that microLED is not a magic bullet that will solve all the issues LCD and OLED have. Instead, each player addresses narrow manufacturing steps and partners with others within the supply chain to meet specific application needs. Display manufacturers have a clear objective in improving LED deposition processes to achieve better uniformity across LED wafers (an opportunity to work with OEMs) and resolve transfer technologies geared toward specific display markets – with wearables likely to be the first offered to consumers.
Unlike LED deposition, which is adopted from semiconductor cookbooks, the transfer step is an opportunity for startups and other innovators to create novel tools and processes for display makers. While the core of the microLED is about display making, the diversity of substrates and form factors, including flexible displays, will have an impact beyond traditional displays and enable product innovations in apparel, fashion, consumer products, and other industries.