Issue



FOUPs, automation, and scale change shape of contamination control


07/01/2004







BY HANK HOGAN

RICHMOND, Va.—In April, Infineon Technologies (www.infineon.com), a subsidiary of Munich, Germany-based Infineon Technologies AG, announced a $1 billion project here that will transform the shell of a fab into a factory capable of processing 25,000 wafer starts per month.

Production of advanced memory chips on 300-mm wafers will begin in early 2005. Infineon isn't embarking on this journey without some real world expertise. Says Henry Becker, Infineon Technologies/Richmond managing director, "More than two years of high-volume production at our Dresden plant have yielded many positive experiences and some important lessons learned for high-volume production on 300-mm."

Although, according to Becker, facility and cleanroom layout are more influenced by the particular product being manufactured than by any other factor, overall trends can be seen in Infineon's approach to contamination control at 300-mm scales. One of the most obvious is airflow.

Turbulent airflow

For years, supplying a steady stream of gently flowing, particulate-free air was the first step in cleanroom contamination control. In the latest fabs, however, airflow is no longer laminar. Thus, the total HEPA filter coverage and the almost imperceptible movement of air found in past generation semiconductor cleanrooms is no more. "It's a turbulent airflow now," remarks Michael O'Halloran, director of technology at Industrial Design & Construction Inc. (IDC, Portland, Ore.; www.idc.ch2m. com). "It's not turbulent on a macro scale, but it's definitely turbulent on a micro scale."

Behind the change is the adoption of Front Opening Unified Pods (FOUPs) to contain and protect product, allowing for more relaxed contamination requirements in the overall cleanroom environment. In the case of Infineon, Becker reports that the facilities have moved from Class 1 (ISO 3) to Class 1000 (ISO 6).

More efficient work space

With the introduction of FOUPs has also come the need for large wafer stockers. Typically, contamination is controlled in the stockers by continuously forcing filtered air through them that washes over the FOUPs and keeps particulates out.

But wafer stockers are not small, commonly starting below the fab floor and sometimes extending beyond the ceiling. In total, it's estimated that stockers consume three to seven percent of valuable cleanroom real estate. As a result, manufacturers are looking for other, more space-efficient, ways to store work-in-progress (WIP).


Infineon Technologies' new Richmond, Va. facility will be equipped with state-of-the-art 300-mm chip production equipment. As semiconductor manufacturing has moved below the 100-nm mark, several new contamination-control implications have arisen???in particular, AMC.
Click here to enlarge image

Says Larry Hennessey, automated material handling system technical manager for IDC, "We're actually placing WIP underneath track sections and allowing the hoist to pick it up and move it to a process tool. So, it eliminates our need to go to a stocker," he explains.

Others, however, are concerned that below-rail shelves, especially those full of FOUPs, will significantly disrupt airflow. Despite the somewhat turbulent nature of the airflow in 300-mm cleanrooms, they suggest a FOUP-filled shelf will likely have dead air pockets below it where contamination could linger. "In general, when you talk about cleanroom airflow, you try to have as few horizontal surfaces as possible," says Becker. "And, if you have to have them, you put holes in them or you have slots for the air to flow through."

No longer toolfarm-centric

Another impact of the use of FOUPs, together with the inherently larger size and weight of 300-mm wafers, is a corresponding increase in the use of automated handling equipment, such as overhead tracks and ro-bots, and a shift away from "tool farm"-centric layouts. In earlier fabs, process equipment was clustered together in "farms" of lithographic, diffusion, etch, metallization and other tools, which allowed certain contamination-control issues to be confined within specific areas. But, as observed by IDC's Hennessey, "today, the toolfarm layout concept has fallen apart because it just takes you too long to go back and forth from point to point to move the lots."


Compared to 200-mm technology, 300-mm silicon wafers contain nearly 2.5 times more chips. As contamination issues become more critical in next-gen technologies, they require point-of-use abatement through filters and other means.
Click here to enlarge image

Instead, says Hennessy, what's emerging is a hybrid approach. For process and contamination-control reasons, lithographic tools and associated equipment still tend to be separated, but other processes that tend to take place together (such as film deposition and subsequent anneal) are now being pulled out of farms and located closer together with implications for contamination control—particularly cross-contamination.


Economic considerations make it unlikely that a fully linear flow will ever be practical. For example, as Infineon's Becker points out, "a wet hood may have ten times the capacity of an etch tool, and teaming these two processes up would lead to severe underutilization of the wet hood."


Other 300-mm contamination-control issues have nothing do with wafer size but rather everything to do with feature size.

As state-of-the-art semiconductor manufacturing has moved below the 100-nm mark, several new contamination-control implications have arisen. In particular, AMC has become a greater concern, with a few well-placed molecules now capable of poisoning a nanometer-feature-size circuit, and requiring point-of-use abatement through filters and other means.

Impact of exotic compounds

The increasing use of what had once been considered exotic compounds is also a growing concern for advanced semiconductor production. For example, the high- and low-dielectric materials being considered to solve size-related electrical problems mean that hafnium and nickel, along with associated compounds that contain them, are now being brought into mass production.

Estimates are that 10 or more new compounds will be introduced into volume production within the next few years, adding complexity to production and waste treatment processes, as well as introducing new potential sources of contamination.