Micro-LED – Quick Take
Micro-LED technology, or the use of ultra-small self-emitting RGB LEDs is a developing concept. Much of what has been publicized by display or LED manufacturers or packagers has been for use in large, modular, and very expensive displays, although Samsung Electronics (005930.KS) has released a 110” micro-LED TV that sells for ~$152,000. Much of the cost has to do with the placement of the 24,883,200 micro-LEDs, which because of their small size (~.002”) must be transferred carefully and quickly from (hopefully) a singulated wafer to a new substrate.
We have indicated in various notes the ways in which companies are trying to make this happen in a cost effective way, and we continue to have on-going conversations with industry participants about the progress on transfer techniques and similar micro-LED issues, but a recent conversation with a potential supplier of micro-LED displays, who is still in the development stage, gives some indication as to how difficult the transfer issue actually is. This potential supplier is able to move 1,000 chips/second which comes to 3.6m per hour. Given that there are 24.88m chips in a 4K TV, at that rate it would take ~6.9 hours to complete the transfer. Considering that a Gen 8.5 LCD fab (15,000 sheets/month) can produce 2 65” large panel displays/minute, it is easy to see why the cost to produce micro-LED TVs is high.
That said, the same conversation also revealed another important metric. The yield on that micro-LED panel is 98.9%, which sounds great, given it is only a bit lower than the 99% yield the company generates on mini-LED products. However when one does the math, a 98.9% yield on a display with 24.883m LEDs means that 273,716 LEDs will not work, and the time and expense of repairing or replacing those micro-LEDs must also be included in the cost calculation. When the display was tested, the non-working LEDs were entered into a ‘map’ that tells the repair tool which micro-LEDs need to be removed and replaced, but instead of being able to make such removals and replacements at 1000 units/second, the tool must find each non-working LED on the map and individually replace it, which happens at a much slower speed, which again has to be taken into consideration when calculating the potential cost of manufacturing micro-LED displays.
While tool vendors and manufacturers ponder such questions and find ways to speed up the process and increase the yield, the display industry marches on, so with even the best manufacturers, despite the announcements and publicity, are still a bit away from competing with existing display technologies with micro-LEDs. It will eventually happen, and will represent a challenge to premium display technologies, whatever they are at the time, but we expect it will take a bit of time for micro-LED technology to be adopted by large panel display manufacturers. “Don’t rush the process. Good things take time” (unknown)
We have indicated in various notes the ways in which companies are trying to make this happen in a cost effective way, and we continue to have on-going conversations with industry participants about the progress on transfer techniques and similar micro-LED issues, but a recent conversation with a potential supplier of micro-LED displays, who is still in the development stage, gives some indication as to how difficult the transfer issue actually is. This potential supplier is able to move 1,000 chips/second which comes to 3.6m per hour. Given that there are 24.88m chips in a 4K TV, at that rate it would take ~6.9 hours to complete the transfer. Considering that a Gen 8.5 LCD fab (15,000 sheets/month) can produce 2 65” large panel displays/minute, it is easy to see why the cost to produce micro-LED TVs is high.
That said, the same conversation also revealed another important metric. The yield on that micro-LED panel is 98.9%, which sounds great, given it is only a bit lower than the 99% yield the company generates on mini-LED products. However when one does the math, a 98.9% yield on a display with 24.883m LEDs means that 273,716 LEDs will not work, and the time and expense of repairing or replacing those micro-LEDs must also be included in the cost calculation. When the display was tested, the non-working LEDs were entered into a ‘map’ that tells the repair tool which micro-LEDs need to be removed and replaced, but instead of being able to make such removals and replacements at 1000 units/second, the tool must find each non-working LED on the map and individually replace it, which happens at a much slower speed, which again has to be taken into consideration when calculating the potential cost of manufacturing micro-LED displays.
While tool vendors and manufacturers ponder such questions and find ways to speed up the process and increase the yield, the display industry marches on, so with even the best manufacturers, despite the announcements and publicity, are still a bit away from competing with existing display technologies with micro-LEDs. It will eventually happen, and will represent a challenge to premium display technologies, whatever they are at the time, but we expect it will take a bit of time for micro-LED technology to be adopted by large panel display manufacturers. “Don’t rush the process. Good things take time” (unknown)
RSS Feed