Keeping pace with Moore’s Law

by Michael O'Halloran, Dennis Grant, Larry Hennessey and Thomas J. Connolly

COMPANIES UNDERTAKING THE TASK OF RETROFITTING 200 MM CLEANROOMS WILL BE CHALLENGED ON MULTIPLE LEVELS OF SYSTEM INTEGRATION.

For the electronics cleanroom industry, the familiar “Moore's Law,” which calls for a doubling of chip performance every 18 months, is the industry's driving force and Achilles Heel.

The continuous need for more product performance relentlessly demands more performance out of every cleanroom that wishes to remain competitive. Ever-increasing cleanroom productivity continues to drive solutions to lower the solutions' costs. That's good news for the consumer, of course, but not always good news for the manufacturers who must commit huge amounts of capital to keep pace with the unremitting march of Moore's Law.

The latest evolutionary stage in that march, 300 mm manufacturing, introduces new technological and economical dynamics that can be either a blessing or a curse to chip makers. The difference between 300 mm's blessing or curse depends on how adeptly manufacturers are able to navigate waters that are largely uncharted.


Putting airborne molecular contamination into perspective.
Click here to enlarge image

300 mm's biggest blessing is its potential to produce higher profits for fab owners. This potential is so compelling that 200 mm factories may approach extinction for volume production of silicon wafers within a year or so. Niche applications will continue to exist in smaller wafer sizes such as Gallium Arsenide, which is just now moving to 150 mm substrates.


Cleanroom scenario without AMC control systems
Click here to enlarge image

As owners consider how and when to embrace 300 mm production, they must consider the economies of the new manufacturing tool set. The most economical production capacity of a semiconductor manufacturing facility is >25,000 wafers per month. This is not a cleanroom or facility issue, it is a tool utilization issue. That yield metric translates into a fab cleanroom size requirement of > 100,000 square feet. Prior to 1995 not many fabs were built at this size. As a result, most 300 mm production fabs will be greenfield.

The demands of 300 mm production have created the perception that it is difficult to retrofit a fab to accommodate 300 mm technology. The validity of this perception depends largely on the individual facility involved. Retrofitting some will be more challenging than others, but with the right approach, retrofitting existing facilities for new 300 mm processes can be achievable, and potentially more economical than building a new fab.

Despite their up-front capital investment requirements, 300 mm fabs are significantly more efficient than previous technologies from a long-term operations standpoint. It is this reality that is compelling all fab owners to contemplate if, when and how they might convert their facilities to 300 mm. This article provides an overview of issues that fab owners need to carefully consider as they analyze the 300 mm “what if” scenarios particular to their specific fabs.


Copper contamination strategies
Click here to enlarge image

Energy
News of rising energy costs and potential shortages is never welcome in the microelectronics industry. The year 2000's sudden energy malaise in the U.S. has re-energized interest in energy among many fab owners.


The ideal fab energy-saving strategy is one that optimizes the inter-relationships among all fab systems, rather than optimizing one system at a time.
Click here to enlarge image

When it comes to energy consumption, 300 mm is not all bad news. Some initial assumptions expected that 300 mm processes would increase power density (watts per square foot) 10 percent or more. While electrical demand for major tools does go up, particularly in the etch area, there is virtually no change in power density per square foot when you factor in the larger footprint of 300 mm tools. That means that while 300 mm fabs consume more energy overall, the greater economies of scale make energy use comparable to 200 mm on a metric basis.

Although wafer fabs are big energy users, energy accounts for only about 4 percent of a wafer's production cost. Energy's modest contribution to the cost of the end product has allowed many fab owners to avoid making energy conservation a priority in the past.

But now looming energy uncertainties are putting more heat on energy as an issue. Energy represents about 80 percent of a fab's utility costs, and there is renewed pressure to reduce that figure.

There has long been a traditional conflict between the sacrifice of “first cost” to implement energy-saving strategies, and the long-term benefits those strategies yield in the form of reduced operating costs over the course of a facility's life cycle. Use of minienvironments in 300 mm fabs has reduced the focus on cleanroom recirculation systems as energy reduction targets. The targets of greatest opportunity for energy savings are exhaust chillers, condenser water systems, and exhaust and makeup air handlers rather than recirculation systems.

Following are some of the energy-reducing strategies that will continue to gain prominence as owners seek to offset the greater overall energy consumption introduced by 300 mm.

Energy savings in each of the following wafer fab systems are often pursued one system at a time, which is a missed opportunity. When owners analyze their fab components as an interdependent whole rather than as independent systems, the collective energy-saving potential rises dramatically.

Key targets for fab energy savings

  • Oil-free air compressor efficiency optimization
  • Humidification air reduction
  • Condenser water heat recovery
  • Lighting & HVAC standards
  • Cleanroom airflow reduction
  • Chiller optimization
  • Boiler autoflame controls

Vibration
Although there are differing opinions on 300 mm's vibration requirements, critical vibration criteria are not dramatically different from those of 200 mm manufacturing. The vibration benchmark for 200 mm submicron production is 250 microinches per second. That standard still applies, with some provisions. The photolithography area is bigger in 300 mm fabs, which can require larger areas of tighter vibration specifications. The 300 mm tools with metrology integrated into the tool may require additional vibration shielding.


Airflow model of a cleanroom before balancing (top) and after balancing (bottom).
Click here to enlarge image

A key vibration variable is the age of the fab being retrofitted. In a retrofit involving an isolated fab floor table without stiffening walls surrounding, there should be careful assessment of structural requirements. In fabs where horizontal vibration hadn't posed a concern in the past, it may now because new 300 mm tools are more sensitive to this type of vibration. Ironically, these large new tools contribute a new source of vibration to the factory floor—vibration of their own making. These added nuances of 300 mm vibration concern can be counteracted with localized structural stiffening, and when necessary, the added support of globalized stiffening.

Another structural issue to deal with is the unprecedented weight of new tools, which has doubled and tripled in certain classes of equipment from 10,000 lbs. or less to 30,000 lbs. or more. In some fabs floor design loads need to increase by 40 percent. In cases where fabs have attempted to skirt structural upgrades by relying on old floors to hold the heavier new tools, floors have cracked, further compromising the fab's vibration integrity and reducing yields. Code upgrades are also driving structural issues as they continue to tighten seismic specifications. The heavier 300 mm tools require bigger anchors and more of them.

Vibration performance testing can be performed on existing structural systems to determine if upgrades are necessary.

Process
Not all 300 mm facilities will involve copper processes. However, those that do need to plan for significant adjustments to deal with copper's ability to spread its contamination throughout a fab.


Outside Airflow Modeling (OSA) also helps reduce cleanroom contamination by revealing how building exhaust and other forms of outside air pollution can enter makeup air systems.
Click here to enlarge image

300 mm tools incorporate modifications to deal with the complications of copper process wastewater. These modifications include separate drains for separate waste effluents—sometimes up to a dozen different waste streams. These more exotic fab effluents require more attention to handling as well, including the need to truck some effluents off site rather than dispose of them in municipal systems.

Fab owners can anticipate a need for more process equipment space, and more power and water consumption for additional process equipment such as slurry and chemical mechanical polishing systems.

Some good news is that the new breed of tools offers improved potential for water reclamation.

Contamination control
Because 300 mm tools are being designed for minienvironments, the cleanroom classification is being degraded from ISO Class 3 (Fed-Std-209E Class 1) to less than ISO Class 5 (Class 100). Per square foot of cleanroom, fewer fans are necessary, along with fewer filters and fewer supplies.

Airborne molecular contamination (AMC) is a growing concern for manufacturers using advanced processes.

The AMC problem is currently concentrated in the more sensitive processes such as memory, but continued reduction in chip geometry dimensions will spread AMC concerns to a broad range of technologies.

Airflow modeling of planned cleanroom configurations early in the design process helps avoid the building of flaws into the room. Modeling reveals how contaminants can migrate to infect adjacent zones in the fab, especially with more volatile contaminants such as copper, and point the way toward mitigation strategies. Modeling also reveals the airflow impact that individual tools introduce into cleanrooms.

The new generation production tools come with guaranteed contamination control performance levels, provided ambient air conditions meet certain specifications. This is not as desirable as it may sound for fab owners, because there are many variables in any cleanroom that can nullify such tool performance “guarantees.”

Click here to enlarge image

The strategic selection of construction materials can be an effective means of taking preemptive action to keep your cleanroom cleaner. Some fab owners are building or having built databases documenting the offgassing characteristics of a host of cleanroom construction materials. Owners who fail to be very selective in their selection of construction materials for cleanroom retrofits may be unwittingly building sources of contamination into their own cleanrooms.


Typical 200 mm tool installation (top); Typical 300 mm tool installation (bottom)
Click here to enlarge image

Filter manufacturers have attempted to keep pace with a new generation of chemical filter solutions for makeup air handlers, recirculation air equipment and minienvironments. The negatives of filtration are that it is an expensive and maintenance-intensive approach to contamination control. The life expectancy of filters is difficult to determine and can vary widely depending on conditions in a given cleanroom. It can also be difficult to detect when filters begin to fail and carbon-based filters pose shedding concerns. One chemical spill in a cleanroom can wipe out an entire set of filters.

Other advanced approaches to contamination control include chambers that scrub air with electrically charged mist, ultraviolet light and water washes. The advancement of AMHS systems is making a major contribution to the long-held goal to reduce the number of people, the largest contamination sources, from fabs.

Tool installation
The migration from 200 mm to 300 mm tools has caused a number of changes to tool hookup, some expected, some unexpected.

As one would expect, tool footprints have increased. Fab level equipment can be as large as 2.8m high, 2.8m long, and 2.3m wide with weights up to 8.4kg. Subfab equipment has also increased in size and quantity, with footprints up to 1.5 times fab level equipment footprints (i.e. for dry-etch areas) and single components up to 2.5m high, 2.0m long, and 1.7m wide, weighing up to 6.0 kg. These increased sizes trigger increased move-in requirements, including:

  • Crated entry points 2.8m wide x 3.0m high.
  • Uncrating rooms 6.0m x 6.8m x 3.0m high (minimum).
  • Wipe down rooms 6.0m x 6.8m x 3.0m high (minimum).
  • Tool move-in paths with 3.0m turning radii.

To facilitate denser tool packing even with the larger fab and subfab footprints, many 300 mm tools have bottom feed connections. While this aids with tool layout density issues, it increases the need for above-slab/below-raised-floor clear space for horizontal piping to allow for bottom feed positioning. The illustrations depict the impact of bottom vs. side feeding utilities on tool packing and underfloor space needs. There are other issues, including capacity sizing, automation and vibration criteria that separate 200 mm from 300 mm tool sets and hence their respective tool installation design and construction.

These include increased electrical demand for major tools, particularly in the etch area (208V greater than 250A), which is driving more 480V supply and distribution to the tools. Electrical interconnects between tools and support equipment are a major consideration in equipment facilities routing between floors.

Regarding tool issues specific to retrofits:

  • Subfab components will still fit in the tool footprints for 300 mm.
  • Consumption of electricity and nitrogen may require an upgrade of supply and distribution systems.
  • Increased electrical demand, particularly in etch areas, also increases the need for cooling and central plant HVAC upgrades.
  • AMHS needs to be fit into existing ceiling systems. Structural and architectural issues need to be solved.
  • The pitch of the bays will change from 3:1 to a 4:1 chase-to-bay ratio.

300 mm manufacturing facilities require the use of expanded automated material handling systems (AMHS) as compared to 200 mm facilities. The key difference occurs in the application of intrabay automation systems designed to move the front-opening unified pod (FOUP) from storage units located within the bay directly to the process tool load port. Hoist systems have been designated as the automation tool of choice for this application.

AMHS is a critical part of a retrofit planning process because AMHS considerations essentially drive the fab layout process. With proper planning, AMHS enhances tool layout flexibility.

Click here to enlarge image

When retrofitting existing 200 mm cleanrooms to support 300 mm manufacturing, the expanded role of the automation system affects the integrity of the cleanroom. In most cases the impact is not a cleanliness issue. The use of FOUPs and minienvironments has significantly reduced the critical aspect of AMHS path routing and potential contamination. Retrofit cleanrooms struggle with the necessary space requirements and loading combinations generated by these systems.


300 mm fab elevation detail (Japanese systems) (top); 300 mm fab elevation detail (PRI system) (right)
Click here to enlarge image

Another challenge to space management of a retrofitted fab is the requirement of higher storage availability within the stocker system. This higher storage volume is required to support WIP bubbles that will be generated during normal operating conditions of the fab. Full automation implies the removal of cleanroom staff from the lot movement task. Thus automation becomes the primary resource for lot movement and must include enough bandwidth to accommodate such situations. 200 mm fabs rarely included this consideration.

An example is a minimum AMHS ceiling height requirement of 4.1 meters. This height is necessary to allow move-in clearance for tools with heights approaching 3 meters. Ceiling heights lower than 4.1 meters do not allow adequate track clearance configurations so tools can pass without track removal.

Depending upon the configuration and ceiling height of the fab, modifications to the cleanroom floor and ceiling may be necessary to allow stocker penetration. Taller stockers create greater opportunity for increased capacity while minimizing the impact to the fab footprint. However, extensions into the ceiling structure create issues with equipment installation and maintenance, and are limited to existing facility clearances above the ceiling. This approach is limited to the configuration of the air handling system in the ceiling and allowable space.

Summary
The wind that powers the microelectronics industry is the development of new applications, which demand ever-increasing performance from semiconductor devices. The encouraging news for those who service the electronics industry is that there are many new applications building momentum today. For the next few years new product directions bode well for strengthening microelectronics demand. This in turn makes the need for a proliferation of cleanroom retrofits inevitable.

The latest advances in retrofitting technology offer new opportunities for owners to extend the lives of their fabs. With careful analysis, there is an increasingly impressive array of strategies available to help owners recover more return on investment from their facilities than they may have considered possible in the past.

Michael O'Halloran is director of technology at IDC where he coordinates IDC's technical initiatives in the rapidly changing high-tech environment and is an active participant in developing the concepts for future semiconductor facilities.

Dennis Grant is a specialist in cleanroom design who has served as lead engineer for the design of some of the world's most advanced 300mm microelectronics manufacturing facilities, including both retrofits and greenfield wafer fab projects.

IDC's AMHS Technology Team is managed by Larry Hennessy, an AMHS specialist who has led many AMHS projects for leading microelectronics manufacturers. He is a specialist in the use of advanced simulation models to test and validate AMHS system designs before they are implemented.

Thomas J. Connolly is manager of advanced technology for IDC, and an engineer with IDC's industrial engineering department.

POST A COMMENT

Easily post a comment below using your Linkedin, Twitter, Google or Facebook account. Comments won't automatically be posted to your social media accounts unless you select to share.