MicroLED Pits Big Apple VS. Tiny LED Chips

　　Differentiating one smartphone from another is no easy feat. A display, however, is the one constant that smartphone vendors believe they can depend on to wow their customers. A new display technology with visible differences in a screen size, resolution, brightness and power consumption could scramble the market.

　　Apple’s anxiously awaited iPhone X, unveiled just this week, is the first iPhone to feature an OLED display — long after competitors Samsung and LG brought to market smartphones with OLED. Of course, unlike Samung and LG, Apple doesn't have its own display technology. Yet.

　　What if Apple were to develop a display technology of its own, featuring all the advantages of OLED but even better … like microLED? While the company has remained tight lipped — as is its custom — Apple has amassed an impressive portfolio of micoLED patents and there is speculation that it is using a Silicon Valley fab it bought from Maxim in 2015 to develop the technology. And Apple isn't the only big name tech company working on microLED.

　　MicroLED displays consist of an array of microscopic LEDs forming individual pixel elements. Unlike OLED, microLED uses conventional gallium-nitride LED technology. MicroLED promises range from high brightness, high dynamic range and a wide color gamut to fast refresh rates, wide viewing angles and lower power consumption. MicroLED proponents claim their total brightness can be 30 times that of OLED products while offering higher efficiency in lux per watt.

　　Whether Apple will trigger the shift to microLED has long been a topic of intense discussion among Apple watchers and display technology experts, not to mention driven speculation that Apple would soon use microLED in Apple Watch. (Much to the chagrin of microLED proponents, the Apple Watch Series 3, unveiled this week, features an OLED display).

　　EE Times this week talked to Eric Virey, senior market and technology analyst at Yole Développement. We asked him to break it all down, and tell us about where he thinks the display market is with this mythical microLED technology.

　　Why do we care?

　　Fully aware of all the hype that has gone into the new display technology, Virey understands what we all want to know: “Does microLED even exist? Why do we care?”

　　Yole does not expect microLED to arrive in the wearable device market, such as Apple Watch, until the end of 2019. Setting the timeline aside, though, Virey believes the electronics industry should pay even more attention today, because microLED’s emergence is going to change not only the face of mobile devices but also the Who’s Who of the display industry.

　　Now Google jumps in

　　The race to secure key patents, manufacturing knowhow and expertise in microLED doesn’t involve the usual suspects — traditional consumer display manufacturers based in Asia. Instead, big guns like Google, Apple and Huawei are driving microLED investment.

　　Most recently, Google entered the microLED fray by investing in Glo, a spin-off of Lund University. Google led Glo’s latest round of investment in August. One Swedish website reported that Glo received $15 million through a rights issue directed entirely to Google. This makes the Mountain View-search engine giant an owner of just over 13 percent of Glo’s capital.

　　Although it did not generate much media coverage beyond Sweden, Google’s investment in Glo reflects the industry’s increasing interest in the new display technology for VR glasses, mobile phones and tablets.

　　Glo today has a team of 40 to 50 engineers working in Silicon Valley, according to Yole’s Virey. Last February, Glo teamed up with Jasper Display Corp. in Taiwan to show full color microLED demo at Photonics West 2017, and later also at Display Week in May, he added.

　　The deal that initially put microLED on the map was Apple’s 2014 acquisition of startup Luxvue, a developer of low-power, microLED-based displays for consumer electronics applications. Apple’s investment in Luxvue has generated a feeding frenzy for rumors and speculation on Apple’s big push for microLED.

　　Not to be outdone by Apple, Facebook’s Oculus Rift virtual-reality business unit also bought, about a year ago, a startup called InfiniLED, spun out of the Irish technology research lab Tyndall National Institute in 2011.

　　Earlier this year, Foxconn, too, revealed its interest in microLED. Teaming with its display subsidiary Sharp, Foxconn acquired a 31.82 percent stake in eLux, a Delaware-based startup engaged in R&D of micro-LED technology focused on virtual reality and augmented reality devices. eLux was founded in October 2016 by researchers formerly employed at Sharp’s research arm in the United States.

　　A host of investment and acquisition activities around microLED startups reflects the tech industry’s never-ending quest for a new generation of display technologies.

　　Challenges all along the supply chain

　　Yole’s Virey noted, “We’ve seen proof of concept, prototypes of microLED displays. The technology is here.” What’s not here, though, is “cost and yield” critical for high-volume manufacturing of microLED.

　　Further, challenges for microLED exist all along the supply chain, he added.

　　But before getting into further discussion on microLED, here’s a primer on differences between LCD, OLED and microLED.

　　LCD is a flat-panel technology that depends on an LED backlight for illumination. The light passes through a matrix of liquid crystal “light switches” and color filters consisting of individual subpixels, Virey explained.

　　In contrast, OLED is an emissive display technology in which each sub-pixel is a tiny light emitter. Brightness can be individually controlled.

　　MicroLED is similarly self-emitting. But it uses individual, small LED chips as its emitters.

　　LED chip production challenge

　　In theory, microLED manufacture should be no different from fabricating LED chips. But regular LED manufacturing facilities don’t fit the job, largely because microLED chips are a lot smaller than regular LED chips. This means every LED chip fab needs a new infrastructure with a “much cleaner clean room,” and “higher resolution lithography,” Virey explained.

　　A better fit would be semiconductor foundries, observed Virey. “It would make sense if Apple teams up with TSMC,” for example, to eliminate the first hurdle of microLED manufacturing.

　　Energy efficiency challenge

　　The production of LED chips causes some insignificant “sidewall” damage, usually one to two micrometers on a regular 250 micrometer x 250 micrometer LED chip. But for a LED chip as small as 5x5 micrometers — necessary for microLED — a two-micrometer sidewall defect would have a devastating impact, leaving a tiny usable area, only 4 percent of the total chip size, explained Virey.

　　To solve such energy-efficiency issues, the microLED industry needs “a two-pronged approach,” the Yole analyst noted. That could involve the development of both new chip designs and manufacturing technology. For example, Glo is working on nanowire. Aledia in Grenoble, spun out of CEA-Leti, is similarly working on microwire technology, observed Virey.

　　Aledia claims it has developed a way to “grow high-density, coaxial gallium nitride (GaN) microwires directly onto large-diameter silicon wafers by using processes that are fully compatible with today's CMOS semiconductor foundries.”

　　Virey sees such an approach — “growing epitaxial layers in a tube” — as a disruptive technology that could bring a breakthrough.

　　Assembly challenges

　　Let’s say we now know how to manufacture tiny LED chips. The next — bigger — problem, said Virey, is the assembly challenges associated with microLED. The question is, “How do you transfer these tiny chips to the back of a display?”

　　To transfer individual LED chips to a 6-inch microLED display would take four days, Virey calculated.  In fact, more than a dozen companies are trying to solve this “chip transfer technology” issue, he said. One way to beat the problem is what Virey calls “the monolithic approach.”

　　The monolithic approach allows microLED to grow directly onto the wafer at the pitch of the final display. “This way, instead of cutting and transferring each individual microLED onto the display backplane, you can cut a large array in the wafer, let’s say up to 1 inch lateral dimensions, and assemble this array (which will contain hundreds of thousands or millions of LEDs) directly onto the driver backplane,” Virey explained.

　　In this case, chances are that this backplane will be made from a traditional silicon CMOS wafer, rather a glass-based TFT, he noted. Conceptually, it's possible to assemble the entire wafer and cut the displays afterward. Alternatively, rather than bonding and interconnecting the array, you can grow the pixel driver circuits directly on top of the MicroLED wafer, he added.  “Lumiode, a New York-based startup, is doing just that,” he said.

　　While the monolithic approach is good for very small displays (“microdisplays”) with very high pixel density, such as >2000 PPI (pixel per inch), this is not a universal answer for microLED.

　　Why? Because the LED wafer size would be limited to 4x6 inches. If the display pitch is too large, most of the precious wafer real estate goes to waste, Virey said. “Imagine 5 um microLED positioned at a 100 um pitch at the surface of the wafer. You’re wasting 99.75 percent of the wafer surface!”

　　For lager displays — necessary for mobile devices, TVs and monitors — most microLED display players are using “pick and place” methodology to put tiny LED chips on the back of the display.

　　Meanwhile, key players in monolithic integration include: Lumiode, Ostendo, Aledia, mLED, Nth Degree and probably Facebook/Oculus, Virey said. Various research organizations such as LETI, ITRI and Hong Kong University of Science and Technology are using the monolithic approach to make microdisplays, which could serve applications such as augmented and mixed reality headsets or micro-projectors such as head up displays.

　　Defect management and repair strategy

　　Among the issues related to manufacturing microLED displays, the industry needs to put in place a defect management and repair strategy, Virey said. At a time when most high-end displays are guaranteed with zero defects, microLED displays will have a tough time competing in terms of PPM (parts per million).

　　Assume that microLED’s yield — which combines both epitaxy and chip manufacturing — is 99.9 percent, as a result of dead or dim microLED pixels. Then the transfer and interconnect yield — that includes dies not properly picked or placed, or faulty connection to the TFT — is also 99.9 percent, due to missing, dead, or “always-on” pixels. Multiply the two yield rates, the combined defect rate of microLED results in 2000 ppm. This certainly isn’t good enough.

　　Numbers like this are triggering massive research to develop microLED testing methods, Virey concluded.

　　Who’s winning?

　　Having followed patent activity in microLED, Virey explained that it isn’t just Apple, Google and Facebook that are working on the technology. Many research institutes, display makers, LED makers, semiconductor companies and startups — “a lot of smart people” — are also heavily involved.

　　Among these players, Virey observed that Apple (after its Luxvue acquisition) “has by far the broadest patent portfolio” in microLED. LG and Huawei are also strong contenders, he added.

　　Sony is also an early developer of microLED technology. The Japanese company, said to be engaged in microLED innovation since 2008, showcased a 55-inch full HD microLED TV prototype at the Consumer Electronics Show in 2012. But since then, Sony hasn’t said much, said Virey. Their focus seems to be more on the industrial/commercial market for big screens, he added.

　　Remember Maxim’s Fab Apple bought in 2015?

　　As simple an idea as microLED display seems (a display made of LED chips), the technology is stymied by manufacturing challenges. Accordingly, many microLED companies are dwelling on problems like assembly and testing.

　　Apple for instance.

　　In December 2015 Apple bought a wafer fab in San Jose, Calif. from analog and mixed-signal chip vendor Maxim. What Apple intends to do with that fab has stirred a lot of speculation. But now we seem to have a fairly credible answer.

　　Apple’s Luxvue appears to have MEMS-based printing technology to place those LEDs precisely onto backplanes. Yole’s Virey believes Apple is developing a specific MEMS-based technology to transfer these tiny LED chips and place them onto backplanes of microLEDs. What used to be Maxim’s small MEMS fab would be perfect for Apple to do test runs, as the company perfects the technology of those “transfer heads.”

　　Yole acknowledges that microLED’s remaining technical and manufacturing challenges could prove too difficult to lick. This is a haunting possibility.

　　Yole sees smartwatches as the “low-hanging fruit” for microLED. Ultimately, though, Yole noted that microLED “won’t completely displace OLED and LCD, but could end up with a strong position in niche applications such as wearable, augmented reality, mixed reality and heads-up display.”