Exposure equipment trends in TFT-LCD manufacturing
03/01/2007
Made manifest by any trip to an electronics store, LCDs are now everywhere in applications from cell phones to televisions. In 2006, the total display industry exceeded $92 billion in revenues and will grow to more than $114 billion in 2010. TFT-LCDs are the dominant display technology and will remain so into the foreseeable future, accounting for almost 79% of the total display market in 2006 and 83% in 2010.
This rapidly expanding industry is being enabled by tremendous worldwide capacity growth from just 4.5 million m2 per year in 2000 to a forecast 128.6 million m2 in 2010, increasing at a 40% CAGR. Capacity growth is the result of heavy investment and the constant shift towards larger and larger glass substrates on which multiple LCD panels are simultaneously fabricated. In 2000, fabs built for 680 × 880mm glass were leading edge; in 2006, substrates almost 9× that size with dimensions of 2160 × 2460mm are state-of-the-art. More than anything else, TFT manufacturing history has been characterized by this rapid growth in substrate size.
Maintaining an evolution
TFT-LCDs are controlled by an active matrix manufactured on nonalkali glass commonly produced with a five-mask lithography based a-Si process. Passive color filters are typically made in a separate, less precise lithography process that may vary from four to six mask steps. In the cell process, these two substrates are aligned, filled with liquid crystal, scribed into individual panels and inspected. In the module process, polarizers, display electronics, backlights, and other components are added to complete assembly before shipment.
The array process is by far the most technically challenging and expensive of the processes. Key a-Si array process tools include PVD, PECVD, exposure, coater/developers, wet/dry etch, wet/dry strip, cleaning, test, repair, and automated optical inspection (AOI). The basic technology related to these machines, relevant trends, market participants, segment forecasts, and market share, as well as information on cell, module, and color filter tools, are all detailed [1].
Lithography is key
Lithography is perhaps the most critical of processes. Typical a-Si TFTs have critical dimensions ~3.5µm, and alignment accuracy of ±1µm is sufficient. Array patterns for large-area applications are mainly exposed by Canon mirror projection or Nikon multilens scanning-type machines. Although there has been some back and forth shifting in market share, the two companies have split the array opportunity nearly 50/50 since 2000.
In color filter manufacturing, only the black matrix step requires <10µm resolution and alignment accuracy <±3µm. Red/green/blue (RGB) pixels, spacers, and vertical alignment (VA) protrusions usually do not necessitate resolution precision <20µm. For these reasons, the black matrix may be patterned with a Canon mirror projection tool, while other features are exposed by proximity printers. Although LCD exposure specifications are very loose compared to semiconductors, the challenges are in exposing larger areas and maintaining throughput. The primary method of maintaining productivity as substrates have grown has been increasing the size of the mask and exposure field. The largest masks used in production today are 1220 × 1400 × 13mm; with a pellicle attached, these can easily cost more than $400,000 for a single mask.
Total TFT-LCD manufacturing equipment and exposure tool revenues by year. (Source: 3Q06 DisplaySearch TFT-LCD equipment forecast) |
Not only are masks expensive, but the total cost of exposure is of substantial concern for panel manufacturers. In the case of array exposure, average machine prices rose 757% between Gen 2 and Gen 8, while the average array tool price increased only 259%. In terms of throughput, Gen 8 exposure is 37% slower than it was at Gen 2, and the average decline in throughput was 19%. So in 2000, the lithography cost accounted for only 14% of total array equipment spending, but in 2007 it is expected to consume 26%. For these reasons, exposure is a prime target of cost-cutting strategies.
The figure shows a forecast of TFT-LCD manufacturing equipment revenues for array, cell, module, and color filter processes on the primary y-axis. On the secondary y-axis, revenues for array and color filter exposure systems are shown. Although exposure costs are going up, on a macro level they track total spending.
Reducing exposure costs
Because there is still no viable mass-production alternative to lithography that can meet the high throughput, resolution, and overlay requirements of the array process, reducing exposure cost is a matter of incremental efficiency evolution. Moving forward, panel makers are expected to increase adoption of four-mask processes that rely on gray-tone lithography to simultaneously pattern both the source/drain and Si island. With TV panel pixels relatively large, even for full HD resolutions, some panel makers are applying proximity exposure for less critical steps. Another array technique is to employ a larger mask on smaller substrates to improve exposure throughput and enable “super” high efficiency lines.
|
Because of less stringent patterning requirements, more of a revolution is taking place in the color filter process, which could eventually completely eliminate the need for lithography (see table). This can be achieved by moving the spacer and VA protrusion steps to the cell process by adopting ink-jet deposited spacer balls and optical alignment. The black matrix can potentially be patterned by multihead laser ablation, and the RGB elements printed by inkjet, offset printing or thermal imaging. Inkjet printing has the most momentum of these methods and is already in production in Japan on large Gen 8 substrates. About 40% of the cost of the color filter is the glass substrate, but even so, by adopting alternatives to lithography, color filter costs can potentially be reduced by up to ~20%.
Conclusion
Reducing lithography costs is an important trend in LCD manufacturing, a part of a holistic strategy to achieve all possible cost reductions to keep LCDs on a continuous downward price curve, driving demand to new heights, where everyone will benefit from the ubiquity of TFT-LCDs.
Reference
- DisplaySearch 2006 TFT-LCD Equipment Report.
Charles Annis received his bachelors from Pomona Colleg and his masters in business law from Ryukoku U. He is VP of flat panel display manufacturing research at DisplaySearch, 9-banchi Shimohatsune-cho, Koyama, Kita-ku, Kyoto, 603-8173, Japan; ph 81/75-491-3846; e-mail [email protected].