|
|
|
Low-K Hits High Gear Spin-on, CVD duke it out in near term SUNNYVALE, AND SANTA CLARA, CALIF.--While low-k dielectric films, if you consider only those with a constant below three, have yet to hit the fab line, the materials and equipment companies eager to capitalize on this enabling technology are drawing battle lines and unveiling equipment sets. In recent weeks, AlliedSignal Electronic Materials (ASEM), a strategic business unit of AlliedSignal, entered phase two of its low k dielectric materials plan with the construction of a multi-million dollar low k dielectric integration facility at its Sunnyvale Calif.-location. Tokyo Electron Limited (TEL) last week introduced its new ACT 8 SOD system for the application and processing of spin-on dielectric materials. The ACT 8 SOD will be launched at a special TEL Low K Dielectric technology seminar on Dec. 4, 1998 in conjunction with the Semicon/Japan '99 trade show. Meanwhile, Santa Clara-based semiconductor equipment giant Applied Materials unveiled its Equipment and Process Integration Center (EPIC) facility and its Black Diamond low-k dielectric film. In contrast to AlliedSignal, Applied's low-k dielectric film follows a chemical vapor deposition (CVD) process. AlliedSignal has been working on dielectrics since the early 1980s, but it was not until the early 1990s that the company saw a need between 0.25-micro to 0.18-micro to move to low-k dielectric materials. The company immediately established a relationship with Sematech for a program exploring low-k dielectric materials. Three low-k dielectric materials, Flare 2.0, Nanoglass, and a hybrid of the two, have emerged from this development work. AlliedSignal sees spin-on as the way to go with low-k dielectrics, a position largely support by industry analysts until Applied Materials' recent Black Diamond product became available. "Certainly Applied Materials is a formidable competitor out there, but there has been a struggle for equipment suppliers to come up with a low-k solution. If you look at their ability to have an extendable solution, there are materials that already have a lower k constant and have been integrated," says David Richter, an AlliedSignal spokesperson. He adds that Applied "has not demonstrated a low-k CVD material that is extendable down to 2.0." Farad Moghadam, VP/GM of emerging dielectric CVD technologies at Applied Materials claims that Black Diamond is sufficient for 0.13 technology and 700MHz. The Applied executive counters that it has already demonstrated the capabilitiy of going down to 2.2. Black Diamond will be introduced as a family of low k dielectric films that are expected to extend to a dielec- continued on page 36 continued from page 34 tric constant of k less than/equals to 2.4. The first product, named BD27, with a dielectric constant of 2.7, was designed for 0.18-micron technology. BD27 will be available initially on Applied Materials' Centura platform in a configuration with up to four process chambers and later on the company's new Producer platform, including future versions of the systems for 300mm wafers. All Black Diamond-series films will be based on the company's DLK chamber hardware to provide multi-generation low k capability. While Applied's detractors may say 'show me the K value,' Moghadam asserts that "K value is only 20 percent of the story, the other 80 percent is how well they can be integrated and how stable they are." In this space, he volleys back integration and stability questions for xerogels and aerogels that can go below 2, "but have to be controlled and contained with other layers." These layers, Moghadam says are added to improve the material strength of xerogels and aerogels, which are basically porous forms of silica. By adding these layers, he says, the K value can rise. "When you do that add a cap you add a capacitor, says Moghadam, "so the capping would compromise the K value from the film." To support its integration efforts, AlliedSignal's low k integration facility will be a Class 10 cleanroom featuring low k spin coaters, curing equipment and film characterization tools. In addition, the cleanroom will be equipped with CMP, etch CVD and strip modules to support the development of integration processes for multi-level structures supplied by leading-edge customers. "The new low k integration facility complements our existing wafer fabrication operation in Columbia, Md., and positions WFG to provide our customers with complete solutions from wafer coating processes through reliability testing of high-performance packaging," said Jack Bolick, VP/GM of the Electronic Materials Wafer Fabrication group. Applied is also doing process integration work at its EPIC facility. According to Moghadam, the company has already shipped to a number of beta-sites, where customers are doing integration work. Perhaps, Moghadam's most convincing argument at least in the near term is that "CVD is a mainstream technology. People in the fab know how to deal with CVD processes. Our concept with Black Diamond is to give them something as close to SiO2 as currently exists, so that the fab customer doesn't have to buy extra tools, can use existing equipment and doesn't have to do anything extraordinary." This is a compelling argument, according to Ron Dornsief, a principal analyst for Dataquest. The existing standard in SiO2 which is CVD film has years of experience behind it and is well entrenched, however, spin ons may have a higher potential. "The idea is to get as close to air as possible, so far the closest thing to getting there is the nanoporous silica, that has the promise of getting as close to one as possible and that is a spin on solution," says Dornseif. None of those films are out of the R&D stages at the present time, though, and all of them have different strengths and weaknesses. Consequently chipmakers are left with a decision to make. "An equipment (spin-on) and materials change, first, then you make a material change (to nanoporous silica), that's a lot of change," says Dornseif. "The other strategy which appears to be a strategy that Applied is going for is to bridge the gap and make only one change in equipment and one change in materials." According to Applied's assertions, Black Diamond and its accompanying process are close to exisiting SiO2 processes. Consequently, chipmakers would not be required to move in a new equipment set and material set, until the nanoporous silica arrives. There would essentially be one change in equipment when the nanoporous silica becomes available. "So it reduces the amount of change so it reduces the amount of risk. That may be a compelling strategy, but it remains to be seen." |