By Paula Doe
With low prices generating strong demand for flat panel computer monitors, makers of the large-size TFT LCD panels are running close to full capacity and moving to add more. But analysts see only modest gains for tool suppliers.
First the good news:
Taiwan’s leading suppliers say they can only fill 80 to 90% of orders, so they’re advertising to hire at least 1,500 more workers and aim to ramp up to ship nearly 80% more panels this year, or 9 million more units. Samsung’s management committee just approved the roughly $580 million (KW 752.9 billion) investment for new fifth-generation line to start mass production in 4Q02. Sharp plans to add capacity by March to run an additional 10,000 glass substrates a month at its Mie line, now running at full 80,000-sheets-per-month capacity. Taiwan’s government has also reportedly just approved Taiwanese companies’ investment in LCD assembly in China, sending a rush of executives off to explore possibilities for new plants there.
And then real mass-market demand for LCD TVs may finally start to build. Sharp President Katsuhiko Machida says his firm expects to sell 1 million LCD TVs in FY02, double its projected 500,000 sales for 2001. Sharp figures the total LCD TV market will triple this year, to reach 1.5 million units.
Now the bad news:
DisplaySearch’s latest update on LCD capex forecasts total spending on equipment for TFT LCD fabrication and assembly will decline 18% this year, to $4.2 billion, only a slight improvement from last year’s 21% drop. But new plants now being planned should result in a 47% jump in tool sales in 2003, or some $6.1 billion. The biggest market this year will be Korea, accounting for 37% of LCD equipment sales, as Samsung and LG Philips equip their 5G plants. Taiwan is expected to buy 29% of tools, Singapore 20%, as the Toshiba-Matsushita joint venture starts up its low-temperature polysilicon display plant there. Though Japan accounted for about half the LCD tool market as recently as last year, it will drop to only 14% this year, as only Sharp has big spending plans.
The market remains awash in too much capacity to make small to medium displays. “There’s a 30 or 40% supply glut now,” says DisplaySearch Senior VP Sam Matsuno. “Applications are currently going up, but not that much.”
And even low-temperature polysilicon technology doesn’t look so exciting anymore. “Many equipment makers want to enter the low temperature poly market with things like ion implant,” says Matsuno, “But that opportunity is not as large as most people think.” DisplaySearch figures low temperature polysilicon displays were only about 3% of the market by total area in 2001, and will increase only to 12% by 2005. When low temperature poly was introduced, people believed it would enable better pitch and aperture ratios, for higher resolution displays.
“But amorphous silicon made a big improvement,” says Matsuno. “So now there is no difference between aperture ratios in notebook or monitor displays.”
The other touted advantage was being able to build the drivers and other chips directly on the glass substrate, but the cost of driver chips has come down faster than the additional cost of low temperature polysilicon production, with its loss in yield.
“For large area displays, low temperature has completely lost its meaning of existence,” says Matsuno. “The only advantage is for low- to medium-sized displays, or for super high resolution.”
He’s not much more optimistic about the other hot future display technology either. DisplaySearch figures organic electroluminescent displays will still account for less than 5% of the display market in 2005.
Things will get interesting in the low temperature poly market next year, when Toppoly starts the first mass production of low temperature polysilicon in Taiwan, and the Matsushita-Toshiba joint venture ramps in Singapore. The two Japanese companies which merged LCD operations plan to start installing equipment 2Q this year, and to start to ramp in October, using large fourth-generation 730 X 920mm substrates.
“Many people, including myself, are wondering what they’re going to do with all those panels,” says Matsuno. He guesses the ultimate aim is to make low-temperature polysilicon backplanes for active matrix organic EL displays for televisions.
Though the LCD equipment market was 20% the size of that for semiconductor tools last year, few companies have managed to make enough from the niche to effectively smooth out the semiconductor market cycles. Still, the leading Japanese suppliers and Applied Materials all take in a nice several hundred million a year in revenues.
Nikon and Canon divide the $400 to $500 million LCD exposure market. Applied Materials’ subsidiary AKT, Anelva, and Ulvac compete in the similarly sized CVD market. TEL and Dainippon Screen offer coater developers, etch, and strip tools. Photodynamics and Robotech dominate in test and optical inspection.
Lack of standards for the LCD equipment may make turning a profit even harder than with ICs. Every LCD maker uses a slightly different glass size, depending on its product size and product mix and ability to get top yield from the process, so any one tool now sells only five to 10 units, 20 at best, when 20 to 30 are needed to absorb design costs and make money, according to Matsuno.
Another challenge is making the vacuum chambers to process those ever bigger glass substrates. LG Philips’ latest plant will start using the industry’s first square-meter-sized glass (1,000 X 1,200mm) in the first or second quarter this year. Philips and Samsung are now thinking about using 1,200 X 1,600mm sheets. And Sharp is considering using even larger sheets — tatami-mat sized 6 X 3-foot pieces of glass. That means huge headaches in making the chamber and maintaining the vacuum. Some say these largest chambers can’t be welded, so they have to be carved out of a single ingot of aluminum or alloy, which limits the diameter. And the current 5G tools are so big they just barely fit into the cargo bay of a 747 – anything bigger will have to be shipped by boat or truck.
So far there’s only one large-area LCD array plant in China. Jilin purchased used equipment from Toshiba but reports are it’s still trying to improve yields and is not yet in commercial production.