The OLED Association, an industry-based organization, provides a forum for the interchange of technical and market information and regularly updates its members on developments.
LG Display Starts Ordering Equipment for Gen 10.5 Fab
LG Display has begun to order equipment for its Gen 10.5 OLED fab in Paju, South Korea. LG is clearly breaking new ground literally and figuratively, as the deposition tools will be one-of-a-kind, which makes early ordering even more critical and the tool ordering process began in early in 2017 by securing a place in the deposition tool queue with letters of intent, and while not legally binding, it likely began the R&D needed by the tool vendor to develop a Gen 10.5 deposition process tool.
The P10 fab will be the largest format OLED fab unless ChinaStar is able to open their Gen 11 fab more quickly. However, ChinaStar is building two Gen 11 fabs and the OLED segment is expected to be part of the second fab. The LGD fab substrate will be 3370 × 2940 mm or 9.9 m2, while the ChinaStar sheets will be 10.12 m2, just 2.2 percent larger. Both fabs will be using IGZO (Indium Gallium Zinc Oxide aka Oxide) backplanes, currently used by LG in their Gen 8.5 OLED fabs. There are a great number of challenges presented to panel producers both by using IGZO for TFT backplanes and handling large sheet sizes, so both fab projects will be learning experiences in their early stages. The phase-I start date for LG’s P10 is currently projected to be Q1 2019 with phase-II beginning relatively quickly. The LGD P10 fab is expected to cost a bit under USD 3 billion, with the construction phase costing ~USD 1.7 billion and have a capacity for 30,000 sheets per month, however there is a chance that the initial LG might mitigate the risk by using 7500 sheets per month for OLED and 7500 sheets for LCD displays. At 100 percent yield, 7500 sheets per month would produce 720,000 65-inch OLED TVs or 540,000 75-inch OLED TVs per year in addition to LG Display’s other OLED TV production lines. LG Display currently produces its OLED TV panels on Gen 8.5 lines, which are able to produce 2.37 million 65-inch OLED TVs and 1.6 million 75-inch OLED TVs so the phase-I lines of P10 will add ~30 percent to LG Display’s OLED TV capacity.
No NAFTA – Samsung to Move the TV Assembly Plant to the United States of America
Samsung Electronics has initiated pre-emptive feasibility study for building a TV production plant in the United States of America. Given that the current administration has teased that it might remove the United States of America from NAFTA, it would seem Samsung, which assembles TVs for the US market in Tijuana, is concerned about the possibility of a 35 percent tariff on TVs imported from Mexico. Whether the potential tariff is for TV sets and components, or just the sets themselves, would determine if Samsung Display would be able to build a profitable panel fab in the United States of America. If it is just for the sets, Samsung could produce panels and modules in South Korea, and assemble them in the United States of America. If panels are also specified, Samsung would still have to pay the tariff on the raw panels being brought to the United States of America. Given that a 55-inch open cell TV panel averages about USD 175 and the average 55-inch basic LCD TV costs USD 380, the US consumer would see that same set rise to USD 513, and that excludes the higher assembly cost of producing in the United States of America relative to Mexico. Samsung is considering Portland, Oregon, South Carolina, and potential sites in Washington state. The threat by the Trump administration is believed to be what motivated Foxconn to start planning for a panel fab in Wisconsin.
UDC Announces Record Quarterly and Annual Results
UDC reported record revenue and net income in Q4 2017 and in 2017. Revenues were
USD 115.9 million up 56 percent in Q4 2017 and USD 335.6 million up 69 percent for the year. The revenue outlook for 2018 was USD 350 million to USD 380 million, a YoY increase of only 4–13 percent. Sid Rosenblatt, CFO claimed that there was USD 15–USD 20 million in material revenue that was pushed forward from Q1 2018 to Q4 2017 and that capacity growth in 2018 would be small, shifting to 2019, which would result in a 50 percent increase in capacity from 2017 to 2019.
The market reacted negatively to the outlook and the stock was down ~USD 20 in after hours trading. Rosenblatt also discussed an accounting change, which would affect when license revenues are recognized. Previously they were recognized one quarter in arrears and now they will be recognized quarterly based on a pro rata of material revenue collected as a percentage of total contract revenue times (using weighted ASP) the total license fee over the life of the contract.
Japan Display Suffers Fourth Consecutive Losing Quarter
Japan Display Inc. has reported a fourth consecutive quarterly net loss, as LCD orders from its largest client, Apple Inc. continued to drop. JDI posted a net loss of 32.6 billion yen for the October–December quarter, versus a net profit of 7.3 billion yen in the year-ago quarter. The company, money-losing for the past 3 years, has been getting half its revenue selling LCDs to Apple. Aiming to offset declining sales with cost cuts, Japan Display plans to streamline production lines and reduce 30 percent of its workforce, resulting in a special loss of 170 billion yen for the current year. The company is having difficulty tapping new investors for capital for its OLED fab.
Innolux Net Revenues for Q4 2017 down by 11.5 Percent YoY
Innolux has reported that Q4 2017 top line and bottom line was down 11.5 percent and 60.6 percent YoY. The company did report net profits of NTD 37 billion or NTD 3.72 per share on revenues of NTD 329.2 billion in 2017. The company’s total shipments of flat panels reached 28.47 million square meters in 2017, up 5.4 percent from a year earlier. Shipments of small- to medium-sized panels totaled 271 million units in 2017, increasing 20.4 percent from a year earlier. Despite a slowdown in the first quarter caused by seasonality and a market reduction in TV sales, and extension growth in China’s capacity, Innolux president Robert Hsiao expects its overall performance for 2018 to be better than that seen in 2017. The company expects:
Large-sized panels to dip 7–9 percent sequentially in the first quarter with their ASP to fall 4–6 percent on average
Small- to medium-sized panel to sink 17–19 percent on quarter with their ASP to drop 1–3 percent
Commercial-use display products and complete TV sets should increase significantly in 2018 as the company has transformed itself into a system solution provider instead of merely a panel producer, according to Hung Mao-Sheng, president of Innolux TV business unit.
Innolux believes the global supply and demand for TV panels will remain stable in 2018, with demand increasing 5–8 percent YoY due to growing TV shipments and an increase in the average size of TVs, while the supply will expand 8–10 percent due mainly to new capacity of 8.6G and 10.5G fabs in China, Hung said. Innolux also expects its shipments of IT panels, particularly notebook applications, to continue to grow in 2018, buoyed by its automated backend production lines even as the global demand drops by 3–5 percent. The company expects its CapEx for 2018 to more than double to NTD 55 billion due to the purchase of a 6G LTPS fab from its parent company Foxconn Electronics (Hon Hai) in late 2017. The plant will be run as an OEM factory for Foxconn with its revenues to account for 5–6 percent of Innolux’s total revenues a year.
AUO Reports 2017 Results Relatively Flat YoY
AUO has reported January revenue of NTD 26.21 billion, up 1.4 percent MoM but down 7.6 percent YoY. Large panel shipments were 9.53 million units, up 1.2 percent MoM and up 12.4 percent YoY, while small panel shipments were 16.65 million, up 2.5 percent MoM and up 39.49 percent YoY. On the surface these look like good numbers as January is typically a down month sequentially for AUO (5 year average – down by 7 percent), but on a YoY basis, AUO’s January is typically up 7.3 percent and this year was down 7.6 percent. With both large and small shipments up substantially YoY and MoM, and sales up modestly MoM, but down YoY the ASP impact was substantial. Based on January panel pricing data (TV average was down by 0.7 percent, AUO’s efforts to meet the China LTPS pricing, while resulting in new sales is driving the gross margins down. AUO also reported 4Q and full year results.
4Q17 – Large panel shipments. QoQ (22.2 million) – down ~5 percent but better than 1Q 5-year average of –8 percent
- YoY flat a bit worse than the 5-year average of +0.3 percent
Small panel shipments. QoQ (43.45 million) – worse than the 5-year QoQ average of +1 percent; up 19.3 percent QoQ decline reflects strong 4Q rather than weak 1Q; new 18:9 demand is driving small panel unit volume
- YoY down 7–8 percent – better than the 5-year average of 1.2 percent
- ASP – QoQ blended ASP down 2–3 percent in USD
- Utilization was high at ~94–96 percent
Outlook. Increasing TV resolution and panel size – expect 1-inch+ size gain for year
- Additional capacity – Gen 8 phase-III to come on line for TV
- Gen 6 LTPS ramping up 19:9 small panel format – company says makes panels 20 percent larger
Negatives. Increasing Chinese capacity – biggest impact in 2019–2020
- General oversupply – AUO says the new norm
- OLED – little investment until proof that it generates additional sales, at which time it will likely be too late
- Micro LED – doing research, but focused on hi-end as more LEDs mean more drivers, means more cost
2018 supply/demand. Demand – expect large panel area demand up 6–8 percent driven by replacement cycle, increasing size and sporting events TV panels demand is forecast to be flat to slightly up by most sources
- 4Q saw China and North America’s demand was down and Europe’s was up – cited strong 11/11 and 12/12 Chinese promotions as positives but overall sell-thru decline less than 3Q; expect 2018 to be similar
- Indian market demand down due to higher consumption tax
Supply – Up in 2018 as noted but next gen fabs take longer to improve yield – 2H will be worse than 1H
Inventories – TV – 1–2 weeks higher than normal
Expectations – 2018. 4K TV market to be 40 percent of overall market – was 30 percent in 2017
- AUO 4K penetration to be greater than the market
- 8K to be 10 percent of something in 2018
- 65-inch market to be up 40 percent (share)
- 70-inch market to be up 50 percent (share)
- 75-inch shipments to double in 2018
AUO has done a good job of differentiating itself from generic panel producers by developing specialized products, which they say represented ~50 percent of 2017 sales. This has helped them keep utilization and gross margins high, but it also means they have to move quickly to stay on the forefront of new products, which all seem to have compressed development cycles. They are not investing in capacity so their market shares have been flat and will soon drop significantly as the new China (Gen 8+ and Gen 10+) fabs reach MP. They have ignored OLEDs and also micro LEDs, are dropping behind technologically. They have benefited from the reduction in large area capacity over the last 2 years, but the industry is moving to an overcapacity situation and the average price/sq m is dropping along with ASPs. The question is how long they can survive and will the Taiwanese government support them in the troubling future?
Challenges for Micro LED Display Manufacturing
In order to precisely transfer Micro LED to a target backplane, the accuracy of equipment for Micro LED manufacturing is required to be less than ±1.5 µm. However, currently, the accuracy of existing transfer equipment (pick and place) is ±34 µm (multi-chip per transfer), while flip chip bonder features with accuracy of ±1.5 µm (single-chip per transfer), which failed to meet the accuracy requirements of Micro LED mass transfer. Generally, traditional LEDs like 3030 LED with a 3000 µm light source can be transferred by SMT equipment, it is available to transfer chip by die bonder when the size of light source reaches 100 µm, the existing equipment for pick and place faces serious challenges of accuracy when the size of light source reduces to 10 µm. Chip bonding and wafer bonding cannot be applied to mass transfer due to low production capacity and high TACT, therefore, currently wafer bonding must be adopted to develop micro LED technology and products featuring low pixel volume with the existing machines, while challenges exist in production capacity and cost, so various thin-film transfer related technologies will be needed in the future. Five thin-film transfer technologies include electrostatic adsorption, Van der Waals force-enabled transfer printing, laser ablation, phase-change transfer, and fluidic assembly. The fluidic assembly is a high-speed assembly approach that achieves high UPH for various product applications, also it can reduce assembly time and cost greatly. Mass transfer is a critical technology for Micro LED manufacturing, transferring micro LED to a target backplane quickly and accurately will be one of the topics that manufacturers need to work on the most as well as on improving UPH and yield.
Micro LED will initially be applied to indoor displays, smart watches, and smart bracelets, due to the high difficulty in transfer technology and various application products with different pixel volumes, some manufacturers conduct R&D with existing wafer bonding equipment and take the application products with low pixel volume as targets to shorten the development period, also some manufacturers directly develop thin-film transfer technology, which required more resource investment and longer development period because the equipment needs to be adjusted and even redesigned, moreover, there will be more manufacturing problems. However, these are very low-revenue products that do not lend themselves to differentiation from flexible OLEDs. Currently, the highest rated volume transfer is ~250,000 LEDs/min, which calculates to TACT 6000 s for 4K and 24,000 min for 8K assuming no LED redundancy was required. To get to a 2 min TACT would require 50 or 120 transfer tools, respectively.
BOE Receives Another Subsidy
The Chengdu High-Tech Electronics Information Industry Development Bureau has given BOE a subsidy of USD 28.53 million for R&D at its new Chengdu Gen 6 fab that began production a few months back. The Chengdu fab has completed the construction and the equipment build out of the first two of three phases, and will have a total capacity of 48,000 sheets when completed. The Chinese trade press called the fab a Samsung killer and China’s first step toward world dominance in the OLED space, which turns out to be sample displays for evaluation purposes. BOE also confirmed the status of their rigid small panel OLED fab in Ordos, Inner Mongolia, which completed the first phase of its construction back in late 2015. The initial phase was for 2000 sheets per month, and the second (of 3) phase is in the ramp up mode. Actual construction of the second phase was completed in 3Q 2017, but is still producing at much lower levels than its expected 26,000 sheets per month. When all three phases are complete, the fab will have a maximum capacity of 56,000 sheets per month, but that capacity seems to have a MP date in 2023.
HC Semitek to Build Micro LEDs and Other Components
Chinese LED maker HC Semitek has signed an investment framework agreement to establish projects of advanced semiconductors and devices in the Yiwu Information and Optoelectronics Hi-Tech Industrial Park. HC Semitek intends to build, develop, and manufacture key components, including LED wafers and chips, sapphire substrates, UV LED, infrared LED, micro LED, MEM sensors, VCSELs, GaN lasers, and GaN-based power electronic devices. HC Semitek has spent RMB 10.8 billion in total on the grand expansion. It will utilize an area of up to 500 acres and the construction is expected to complete in 2025. The company said that the establishment is in accordance with its strategic planning for future growth and does not affect the Chinese firm’s financial performance in 2018 but will benefit its long-term growth.
VR on TV – Still a Work in Progress
The PyeongChang 2018 Winter Olympics could be watched in virtual reality on several headsets via NBC’s app. Sadly, the VR programming with one exception was spotty and disappointing. During the Rio Games in 2016, NBC broadcast the Olympic opening ceremonies in virtual reality for the first time, and had a solid start – despite problems like poor resolution. 2 years is a long time for technology like VR, so when NBC and Intel announced they would cover the opening ceremonies and several live events once again in 2018, it created high expectations. The games in VR in the United States of America are relatively smooth, provided one has a live TV/cable option and the right headset. Among the models that will work are the Samsung Gear VR, Google Daydream, and Windows Mixed Reality headsets. Simply install NBC’s app at the Oculus, Microsoft, or Google Play store and enter the name of the cable provider, along with a username and password, which grants access to live coverage of several events per day, along with full replays of those events as well as the opening ceremonies and highlights of others. Some broadcasts are available in 180° 3D VR and others in 180° 2D.
The app is also glitchy and one is often greeted with a black screen when loading sports. The only way to fix it was to quit the app and go back in. Both the downhill and giant slalom are ill-suited to VR, at least the way NBC handled it. The skiers zoom by the fixed camera positions in blink-or-you-will-miss-them blurs, so the only real way to follow it is via the VR Cast mode, with cameras controlled by a director. It does give one an idea of what it is like to be a spectator. But other than for a couple of seconds per run, one is watching the same 2D broadcast feed as anyone. NBC and Intel have not improved the image quality much since the last Olympics. It is not bad for close-up shots, but with wide angles it is too pixelated at times to even make out someone’s face. NextVR’s NBA and concert broadcasts, done with multiple 6K RED 3D camera rigs, result in a much better resolution for viewers. The sole bright moment was in curling, because of the relatively calm pace and controlled atmosphere. Most of the cameras are not too wide, so resolution problems were minimal. And curling is a game of depth, so the 180-degree 3D view was ideal.
On the other hand, ice skating was abysmal in NBC and Intel’s VR because of the poorly placed cameras and the necessarily wide view. The latter made it hard at times to even tell if there was a skater on the rink. Ski jumping, snowboard half pipe, bobsled, and luge were also not great, because the athletes move by the cameras too quickly to see much. Again, the frame rates and resolution cannot keep up, so sometimes the athletes are literally just a blur.
As for why NBC has not gotten better at this yet, a big part of the problem could be Intel’s True VR technology. Both NextVR and Intel cover the NBA, for instance, but on a Gear VR headset, Intel’s tech is clearly inferior. It is also obvious that VR is still an afterthought for NBC, letting the network tick that box and save its best for the main broadcast. Here are a few suggestions:
- Do whatever it takes to make the images much better, including ditching Intel if it cannot deliver
- Do not waste time with 360-degree video; it is useless for sports. Focus on 180-degree 3D, and do it better
- Get more creative with camera positioning. VR is useless unless the camera is right in the middle of the action. Bobsled, for instance, would be great with cameras mounted on every sled
- Improve the app so it is more immersive and less glitchy.
The 2020 Summer Olympics will be hosted in Tokyo, and Japan is expected to put on a technological tour de force, with 8K broadcasting and more.