Exclusive: Unilumin talks about improvements to mini-LED manufacturing
At ISE 2019, Unilumin brought mini-LED to the show. What does this mean for the LED manufacturer and what impact will this have on the LED market? Inavate APAC had an exclusive chat with Steve Scorse, vice president of Unilumin Group, to find out.
Scorse agrees that the technology is not something new: “Mini-LED, it’s a four-in-one SMD. So, for a 0.9mm pixel pitch, you have roughly an 8mm dye with red, green and blue in it. Basically, you have four large RGB dyes and with a big lens over the top and we call it mini-LED.”
The technology has been around for a while and Scorse discusses its position in the general scheme of LEDs: “Mini-LED and micro-LED are effectively the same thing. The only difference between them is one of scale. The chip size is what differentiates mini-LED and micro-LED. Mini-LED is between 30 microns and 100 microns. Micro-LED is between 10 microns and 30 microns.”
But there are key differences between the two as Scorse says: “Micro-LED is designed primarily for PDAs and for incorporation into personal devices like smartphones. So, some companies, like Samsung and Sony, are showcasing some micro-LED displays which at the moment are not in what I call mass production. Yes, we can make one. But they’re very expensive and it will take you six months. Micro-LED is not in volume production. The manufacturing techniques are not yet mainstream enough.”
These facts are pertinent, because while Unilumin’s mini-LED might not be new, the manufacturing process behind them is a game-changer. Scorse says: “In the interim, while manufacturing processes for micro-LEDs are being improved, mini-LEDs are out there, and they have as many advantages as micro-LED but with a lot of other benefits as well.
“It’s much more energy-efficient. It’s easier to manufacture and we have already developed a manufacturing process. The key thing to the process is the way we have dealt with pick-and-place machines. Unfortunately, pick-and-place machines can’t work with this because they can’t do enough dyes, the dyes are too small. So, at Unilumin, what we’ve come up with a process called ‘mass transfer printing’. Basically, we have all the dyes on a carrier and then you start the transfer onto a PCP. It is called a flip chip technique and basically the LED comes out upside down.”
The importance of the proper manufacturing process can be seen in the failure to launch of Chip on Board (COB). Scorse says: “One of the reason COB failed is that it required wire bonding. So, you have these red, green and blue dyes and you have the cost of the gold and then you have the reflectivity of the gold because that’s being illuminated by the dye, causing a problem, and then you have the manufacturing complexity. Now, I’ve got to solder wire bonds on every one of these COB tiles. I’ve made manufacturing far more complicated.
“The other problem is now I can’t get to really small pixel pitch because I’ve got this part [the wires] in the way making it twice as big as it needs to be. These are the reasons COB has basically gone away. It’s fine in the middle pixel pitches, 2.5mm, things like that. It works fine. But in a very narrow pixel pitch, it’s not going to work. You end up with poor black level because the gold reflects, and the black is muddy.”
He continues: “With our process for mini-LED now, we can transfer 20,000 dyes at a time. We then put the same sort of encapsulation layer as you would do in a COB. There’s no lenses. There’s no optimal layer. Mini-LED has a very uniform distribution of light. You have a 175-degree viewing angle, there’s no brightness degradation as you go offline because there is no lensing that focuses light at any direction. And now the product becomes waterproof. It becomes touch-resistant. It becomes impact resistance because of the encapsulation that you guys have.”
Scorse says: “Unilumin, we have complete control of the supply chain from the Sapphire wafer to the cutting, to the mass transfer printing, to the surface layer, to the cabinets, to – basically we can control everything. We have invested over USD200 million in a new production process plan for R&D and manufacturing of mini LED.”
The next step is to improve the process. Scorse details: “Your yield has got to be very high. You have to make sure that you’ve got no more than one in 50,000 failures. At the moment, we have 99.9997% yield. But that’s for an individual dye. That means for a videowall, like the one we had at ISE 2019 which had more than eight million pixels, there are 24 million dyes in that display. If I now have a 99.997% yield, that means three failures in a million. Now I suddenly have 72 failures in my wall. That’s not a viable solution. We want to get to 99.999% yield, one part per million failure rate and I think it’s achievable. At the moment, there’s a price premium for this over SMD. But the price premium has all to do with getting the production process optimised for the volumes.”
Improvements in yield and the component cost being lowered should make LED videowalls in narrow pixels much more attractive for a range of applications.
Scorse says: “There’s a lot of applications where a sub-one-millimetre pixel pitch would be ideal. Today, they’re not used for a number of reasons, maybe because the environment is not conducive to sort of having someone accidentally kick that wall and cause USD20,000 worth of damage. So, people are very cautious about these applications where it’s very easy to touch the LED videowall. Part of that is retail and public spaces. But if you can get the cost of mini-LED lowered, the reliability up and the ruggedness up, then you will see this really take over in the control room market for instance. Then you have applications like digitalisation, simulation where small pixel pitches below 0.9mm would really work, these applications would begin to open up in the market.”