Chris Chinnock, Insight Media

This seemed to be the clear message from visits with microLED exhibitors, symposium papers and a panel discussion held on the Center Stage at the recent SID Display Week event. Below is a summary of some of the key comments from the panel discussion. Speakers included:

• Eric Virey, Principal Display Market and Technologies Analyst and Consultant, Yole Group, moderator
• Wei-Lung Liau, CTO, AUO Corporation
• Alex Chen, CTO, BOE
• Charles Li, CEO, PlayNitride

First, microLED chip-maker and mass-transfer provider PlayNitride noted they had achieved their first profitable year. He and the other panelists stated that they thought the microLED industry was better off this year than last year, which was very difficult with the cancellation of Apple’s wearable microLED project. New applications, like automotive are emerging and AUO is in mass production of a smartwatch for a customer, plus supplying panels to Samsung for their TV and Sony-Honda for the Afeela automobile.

However, all the speakers acknowledged that cost is probably the biggest current challenge.  But all see slightly different priorities to driving down cost.  AUO’s Liau for example, sees chip costs (PlayNitride is their major supplier) as the number one driver with mass transfer as number two. BOE’s Chen agrees on chip cost as the most important cost driver with the second being defective detection and repair. The microLED cost is still over 50% of the display cost.

Getting the cost of the microLED device down is a mixture of reducing device size, yield and scale. Chen also said that volumes are still small but the development of device size and pixel pitch standards would clearly help to advance the industry and reduce cost. Both BOE and PlayNitride are fabricating blue and green microLED on smaller 4” or 6” sapphire substrates. Chen said the quality of the 6” is better than the 4”, but it will take more time to cost reduce.

PlayNitride’s Li told attendees that his biggest concern is customer confidence and their ability to commit.  These product makers believe in the microdisplay opportunity, but they are not sure when it will arrive. PlayNitride has established the capacity and reasonable cost structure to move forward, he believes, “so don’t worry about capacity and cost.”

Virey then noted that microLED chip size seems to have plateaued at 15×30 or 20×40 micron device sizes, so how can the industry move to smaller devices? In response, Liau noted that as devices get smaller, in addition to the yield loss, there can be losses in efficiency, which means power consumption would go up to maintain the same luminance levels. This can cause downstream application issues. With smaller devices, mass transfer yields will also decline. “Neither is a fatal issue,” noted Liau, “we have ways to address this so it is a timing issue for us.”

Chen thinks that the 15×30 micron size is best for mass transfer now, but they would consider moving to 10×20 micron flip chip and longer term, a vertical microLED (stacked microLED) is a good direction, but it needs more research, especially the mass transfer repairing concern.

Li expressed a bit of frustration with the industry as customers tend to be discussing options too long and not locking down specifications in terms of the microLED specs, TFT technology, optics, transfer equipment, etc. Potentially, there is no limit for device size reduction – we just need to lock down a size to focus on the qualification of devices to move projects forward. It is not urgent to reduce device size right now, he thinks, all aspects of manufacture need to be reduced simultaneously.

The conversation then turned to mass production substrate size for mass transfer. Here, many developers have been using Gen 2.5 sized substrates to do the mass transfer, but AUO has announced they are building a Gen 4.5 line. This enabling the production of 42-inch microLED modules – the biggest in the industry. To create a 127-inch home cinema screen requires only 8 panels, in this case. The speakers acknowledged that even larger-sized substrates will help to reduce costs, but the industry is not ready to move to Gen 6 or Gen 8.5 yet. For one, mass transfer is not mature.

BOE’s Chen believe Gen 6 may be a good sweet- spot as LTPS offers the best driving circuit technology right now but agrees that mass transfer at this size is not available. For Gen 8.5 both AUO and BOE would want to move to an Oxide backplane, but the mobility would need to improve from 20-30 today to greater than 50 to reduce power consumption. Oxide backplanes would be lower cost, if the mobility can be improved. They see Oxide devices fitting mid-tier microLED applications as they would offer lower resolution and brightness than LTPS-based devices. LTPS is still the best today for higher-end applications like higher brightness, resolution, transparency or frame rate.  microIC type solutions would be even more expensive, thinks Liau.

Li offered a novel idea: make the backplanes on Gen 6 lines, then cut the glass to Gen 4.5 sizes for mass transfer. He also thinks there is potential to be able to make transistors in GaN as well as microLEDs.

In terms of second-generation mass transfer equipment, they all agreed they would like to see equipment makers continue to invest aggressively in this area, but the industry needs to show continual progress in terms of volumes and market penetration. Chen noted that a lot of focus is on OLED equipment this, so hopefully more attention will be paid to microLED soon. And, some of these suppliers come from the semiconductor world, so they need to learn how to make equipment for the much larger glass substrates. “We need customers and equipment makers to buy-into this market and invest to move it forward,” said Li.

Virey then asked about production bottlenecks. AUO’s Liau said that yields are going up but microLED chip yield is still too low with too many defects. They would like to see a much faster solution for electroluminescent testing of each microLED instead of the sampling approach they need to implement now.

BOE seems more focused on the mass transfer issue as Chen would like to see better chip-on-carrier technology, along with test & repair equipment after mass transfer. “Each piece of equipment in the process needs tuning to balance the line so we would like to see better standardization here,” noted Chen.

Li says their microLED chip yield has advanced from 85% to 95% to 99.5% now. But they need better and faster inspection equipment at a reasonable cost. He thinks this bottleneck will be resolved in 1-2 years.

In terms of the manufacturing process, AUO thinks it is too early to know the best supply chain configuration right now. Some wonder how much design and/or manufacturing should be in-house or done with partners. Others are thinking about a microLED foundry model similar to that used in the semiconductor industry, for example.

“Right now, we support an Integrated Design Manufacture model where the LED design is done in-house as it must be done in conjunction with the TFT design,” said Liau, “and eventually with a mass transfer campus including logistics.”

BOE wants to work closely with the microLED chip designers at this stage. He thinks each developer should keep doing what they are doing as that is their expertise. This will lead to device efficiency improvements and new microLED fabs. But developing standard sizes and specs is just as important to drive down cost.

As usual, Li asked attendees to think a little outside the box. “Why aren’t microLEDs being considered for lighting? Why aren’t developers trying to leverage the vast infrastructure, manufacturing technology, and markets for conventionally-sized LEDs?

In terms of other approaches like microLEDs-on-silicon, stacked microLEDs, multi-color microLEDs, quantum dot conversion and nanowire microLEDs, PlayNitride’s Li says it is too early to consider these options. Liau agreed and was concerned about mass transfer and power use. Chen and Li think the color conversion option is the most promising for monolithic microLEDs for AR/VR. Chen is concerned about the reliability of the quantum dots and optical crosstalk issues. Long term, the stacking approach may be the best solution.

Moving on to market applications, Liau says that microLED displays should not look to replace LCD or OLED displays but find new applications where their benefits offer a clear value proposition. This includes transparent and stretchable display. While microLED TVs offer a clear value, their cost makes them a niche product right now.

Chen sees consumer wearables as a bright spot, but power use may be holding back developers right now. He also likes the large-sized application, particularly commercial display, with TV later.

Wrapping up the discussion, Liau said that he is optimistic about the microLED industry and invites more partners to join them to help move the industry forward. Chen think microLED more closely resembles semi-conductor manufacturing offering an interesting pathway forward. And finally, Li says that PlayNitride is mass-production ready now, so come talk to them.