"It's an interesting story," said Paul Semenza, manager of strategic-market analysis at display-market watcher Stanford Resources Inc. (San Jose). "People had written off amorphous silicon, but then came the old incremental improvements and other tricks."

Amorphous-silicon (A-Si) LCDs have pushed beyond the 100 color pixel/inch density of the typical CRT and, in a few cases, crashed through the 200-pixel/inch frontier, a level at which observers say the display approximates the quality of print and begins to approach that of photographic film.

Though it is now clear these new levels of resolution can be reached, the question of where that capability will be necessary or in high demand is still being debated. Large high-res screens would be useful for applications such as video editing, graphic design or electronic publishing, while small screens would be ideal for portable consumer products or instrument screens. And if the right price point could be found, large high-res screens could make their way into mainstream markets such as flat-panel desktop monitors.

A-Si LCDs above 100 pixels/inch have recently become available from a number of display makers, including Fujitsu, IBM, LG Philips, NEC, Samsung and Sharp. Ultrahigh-resolution models at 200 pixels/inch or more are in the offing from both IBM and NEC.

Semenza noted further that Samsung has showed a prototype of a 6.6-inch active-matrix (AM) LCD with 300-pixel/inch density, intended for electronic-book applications. Packing an XGA, 1,024 x 768-pixel format into a 12.1-inch-diagonal display requires resolution of 100 or more pixels/inch, while accommodating a quad-XGA, 2,048 x 1,536-pixel format requires more than 200 pixels/inch for the same size.

The pioneer in high-resolution A-Si active-matrix LCDs, dpiX (Palo Alto, Calif.), failed to make a commercial success in the 1990s, despite developing a 141-pixel/inch model in a field of 75- to 90-pixel/inch competitors. But then IBM Corp., at work on AM LCDs since the mid '80s, picked up the challenge. In 1998, IBM Research announced that its Advanced Display Lab (Yorktown Heights, N.Y.) had developed a prototype AM LCD called Roentgen that delivered 200 pixels/inch.

"An early focus of this work was developing techniques to control yield loss in AM LCDs," said Robert Wisnieff, manager of the lab. "The results from this work provided IBM scientists with the [ability] to create highly complex displays" — a development that led IBM to proclaim, at the Roentgen announcement, that "the much predicted obsolescence of paper may be at hand."

The Roentgen technology has become commercially available within the last few months, said Alan Jones, business-development manager for IBM UK (Winchester Hants, England), and according to IBM, its image is "virtually indistinguishable from a printed page." The lab is now working on a follow-on generation, dubbed Bertha.

Moving to higher resolution presents a number of challenges to AM LCDs, as does moving to higher pixel counts and larger sizes. Not least of these is the need for smaller electronic structures with better performance characteristics. "The conductivity of bus bars has been a traditional issue, and maximizing aperture ratio so you have a very efficient and bright panel is the key issue," said Joel Pollack, vice president at the display operation of Sharp Electronics (Camas, Wash.).

Wider aperture needed

"You need a wider aperture in the liquid-crystal cell, otherwise the panel is very dark or we have very high power consumption," said Mark Akiyoshi, senior engineering manager for display products at NEC. "It requires a narrower TFT [thin-film transistor] structure and wiring patterns, and we also need to reduce the black matrix stripes. That means the tolerance between the TFT and color filter stripe has to be more precise."

According to Pollack, the need for smaller features and higher performance can be at cross purposes. "You have to get signals down the bus bars but hide them so they don't kill your aperture ratio, so you have to make them thinner and thinner and still not lose the signal," he said. "Now as they get thinner and thinner, geometry has an effect. It's so many ohms per square of sheet resistance that go down line and, if [lines] are thinner, there are more squares and more signal gets dropped."

Higher-resolution displays require lower-resistivity thin-film wiring, said Fumiei Hayashiguchi, marketing manager at the Display Business Unit of IBM Japan Ltd. He said that quad-XGA and higher resolution levels require a move to aluminum and, ultimately, copper alloy, at 2 to 4 micro-ohms/centimeter.

IBM's original Roentgen prototype, a quad-SXGA, 2,560 x 2,048-pixel display, packed more than 5 million pixels into a 16.3-inch-diagonal screen with 1.64 miles of aluminum-alloy lines. With the Bertha generation, IBM has further improved its wiring but "has not realized copper yet," Hayashiguchi said.

Nor has NEC found its way beyond aluminum, said Akiyoshi, who lamented that "aluminum is not a good material" for the lines of a display. He pointed to an adhesion problem in particular, noting that it's time-consuming and expensive to compensate for this. "This is an area that has to be improved," Akiyoshi said, "and we're investigating [it], but there's no alternative right now."

Akiyoshi noted several other challenges to achieving high-res displays, including the interconnection between the panel and its driver chips. "As the pixel pitch decreases and the patterns become narrower, the connection becomes more difficult," he said. "It's a new challenge for everybody, including the TAB [tape-automated bonding] driver manufactures and the tape manufacturers."

"Clearly, if someone tries to make a UXGA 6.4-inch LCD — God knows what you'd need it for — that could challenge the interconnect technology," said Sharp's Pollack. "To my knowledge, those making even SVGA [800 x 600-pixel] 8.4-inch displays are not finding homes for them."

Moreover, said Akiyoshi, as resolutions rise, "the timing to drive the cell is becoming shorter and shorter, so we may not have enough time to drive the crystal." According to IBM UK's Jones, a quad-SXGA, 200-pixel/inch display needs a 1-Gbyte/second data input rate.

Lack timing know-how

Akiyoshi said that standardization efforts are in progress to "reduce blanking times so we can have a longer driving time for the LCD," but these have yet to be completed. Also, he added, "Many chip manufacturers don't know how to do the timing for high-resolution flat displays, and you can't get [graphics] boards [for hi-res displays] yet. The infrastructure is not ready to support QXGA and higher."

IBM, however, developed what it calls "a scalable graphics adapter architecture" along with Roentgen. Based on off-the-shelf components, it's said to be capable of handling these types of high-image-content displays.

As NEC Electronics prepares for mass production of its 202-pixel/inch technology, first announced in late 1999, the IBM lab is now crafting the 205-pixel/inch Bertha. This is a wide-format 22-incher with 9.2 million pixels, formatted as 3,840 x 2,400 pixels, or "wide" quad-UXGA.

Hayashiguchi noted a number of "technology highlights" in Bertha, including post-spacer technology, fine color filters and a newly developed liquid-crystal material that enables a wide 170° viewing angle and rapid 16.6-millisecond response. Its relatively high, 28 percent aperture ratio, an improvement of 0.3 percent over Roentgen, is the result of primarily short-channel-length TFTs and the company's high-resolution process (HRP) structure.

Bertha makes use of a combination of in-plane switching (IPS) and a dual-domain architecture, as does a 123-pixel/inch, 20.8-inch, quad-XGA AM LCD the company is producing now. The combination boosts viewing angle from 160° to 170° over IBM's previous technology, and it results in a larger aperture than multidomain vertically aligned nematic LCDs from Fujitsu and Samsung, according to Hayashiguchi.

Simpler switching mode

Conventional twisted-nematic LCDs follow a helical switching path, while IPS LCDs use a simpler horizontal switching mode. The benefit of IPS is a great improvement in viewing angle; its downside has been relatively long switching times and reduced contrast. Multidomain LCDs define several switching domains in the liquid crystals within a pixel area, combining the different viewing-angle characteristics of different domains to achieve a wide viewing cone. The problem here has been manufacturing complication, requiring additional structures to properly align the liquid crystals.

The vertically aligned nematic LCDs need "protrusions and/or ITO slits to fix their pretilt direction," said Hayashiguchi, and their aperture shrinks as resolution rises. IPS, in contrast, "has no limitation for initial alignment because of homogenous alignment without pretilt angle," he said, claiming that the combination of IPS and HRP delivers high brightness, low power, wide angle and excellent heat management.

As for Bertha's post-spacer technology, it was put to use earlier in an 18.1-inch SXGA panel from IBM. "In a conventional structure, light leakage around spherical spacers causes contrast deterioration," Hayashiguchi explained. "Post-spacer technology [provides] a more rigid panel construction, preventing light leakage to provide better contrast ratio."

The new HRP structure packs more into less by modifying the traditional AM LCD structure. In the conventional structure, lines lie adjacent to ITO electrodes, with a certain amount of distance required between them. HRP embeds the lines in an organic insulator and fashions the electrodes on top of that insulator, narrowing the structure and increasing the aperture.

Bertha has yet one more technology in its bag of tricks, but one that Hayashiguchi said has not made it to production yet. It's a "one-drop fill" technique for filling the display with liquid-crystal material, with the promise of shaving time and steps off the production process. In combination with post spacers, the technique, now in trials, provides precision control over the LCD's 3-micron cell gap, said Hayashiguchi. "The most difficult part is to cure the seal very quickly."

Akiyoshi noted several items in NEC's arsenal that it applied to a 6.3-inch, XGA AM LCD. The display is due to go to mass production in the second quarter, said Cathy Dotson, strategic-product marketing engineer at NEC Electronics (Santa Clara, Calif.). While it makes due with only evolutionary improvements in most of its basic technologies, the critical hi-res problem of color filter alignment called forth an unconventional approach.

Color AM LCDs typically locate their active matrix and color filters on opposing substrates, while NEC's 6.3-incher puts both on a single substrate. It removes "the worry about proximity of the color filter, black matrix and TFT," Akiyoshi said. "It's a different process on the TFT substrate glass and different layered structures."

How much further can amorphous silicon stretch into the future? "Maybe up to 220 or 230 [pixels/inch], amorphous technology will still be applicable," Akiyoshi said.