Time for a communal 300mm CMP apps lab
11/01/2002
IC manufacturers made it clear at the CMP-MIC meeting in February 2002 that the high cost of CMP slurry in the fab and the desire to have multiple viable slurry suppliers were issues that needed to be tackled. Subsequent discussion between panel members and the audience pointed out that the fab-imposed price of admission for all slurry suppliers is to have a credible applications lab, which requires a considerable capital investment and is recoverable only through slurry prices. It was clear that a novel approach is required to foster a robust supplier community while sheltering slurry costs from uncontrolled escalation.
The 300mm development facility under construction at Albany NanoTech, under the auspices of the Governor George E. Pataki Center of Excellence in Nanoelectronics, provides an opportunity for establishing a CMP development facility that could address this industry-wide problem. The core concept is to create a CMP lab within the facility that houses each of the major CMP polishers, and others, as space and funding permit. Suppliers providing equipment must offer maintenance support and upgrades as necessary for the duration of the program. The lab must also include all of the requisite metrology for wafer processing and slurry and wafer characterization.
The proposed center — let's call it the CMP 300mm polishing operation, or C3PO — would make a wide range of 300mm CMP capabilities available to support other activities taking place in numerous programs already committed. Earlier proposals for stand-alone facilities lacked this essential element. Equipment and consumables would be provided for a fee that supports the infrastructure, which would include a full-time staff to maintain and operate the equipment as necessary, and to train and certify outside users buying time in the lab. Use charges for internal programs could vary according to external revenues; in good times, eligible internal programs could use the facility at no charge.
Certainly large market share and small market share suppliers would have different needs and interests. For large market share equipment suppliers, the lab is an opportunity for more current or potential customers to evaluate new consumables, new process conditions, new applications, and new performance benchmarks on a specific machine — all of which will ultimately help the tool company sell even more machines.
For small-share equipment suppliers, what better opportunity could there be for competitive processes to be developed? Instead of the entire sales, engineering, and marketing forces being consumed by half-interested semiconductor customers repeatedly droning "show me the data," a swarm of graduate students and other paying customers may be inventing the future on their machines. The processes will have been developed with other peoples' resources.
Companies in the small-share CMP consumables category are in luck — after all, this whole proposal is being made to lower their barrier to entry and to reduce the costs that will have to be passed along to customers. This carries the burden of keeping a supplier's costs contained; defining "containment" is between the vendor and its customers.
Large-share CMP consumables suppliers may find themselves in a real dilemma, however. Their own 300mm apps labs are already depreciating and factored into product costs. In many categories, the products of the large-share suppliers have become de facto industry standards, accessible to everyone who practices CMP at any level. As the market share leaders, they have nowhere to go but down if C3PO successfully meets its objectives. If C3PO does open, they will find themselves in the position of making pads available for the slurry vendors to benchmark their latest, greatest slurry offerings. Likewise, their most commonly used slurries will be expected to be available for the start-up pad vendors to benchmark their performance. Either way, the potential competition grows and the market share destabilizes for the large-share suppliers.
However, should a supplier decline to make its materials available to C3PO, it will find itself at odds with its largest customers: the ones who spoke up at CMP-MIC. It's not entirely uncommon for a sympathetic customer to donate consumables to labs working to satisfy its interests. Even more disconcerting should be the potential of small-share consumables suppliers pairing their own products together to create the next killer app, excluding large-share suppliers at any level.
There is some amount of metrology that is absolutely essential to CMP process development and characterization. As devices shrink and specs tighten, the metrology equipment gets more sophisticated and more expensive. Here, too, there is the market leader and everyone else. There is a limited amount of space in C3PO and money to outfit it. It is probably to be expected that the greater motivation to install metrology tools at substantial discounts with prices asymptotically approaching zero will come from those aspiring to obtain market share, rather than those merely wishing to maintain it. It will be up to C3PO's managing staff to determine which equipment best meets the community's needs at an affordable price.
Fabs would expect that one of the benefits of C3PO is a deceleration in the rise in CMP consumables costs, at least a leveling off, or perhaps even a decrease. While intimidating, the cost of installing and operating a CMP applications lab at each consumables vendor is only part of the development cost. Realistically, given the increasing complexity of materials sets and process integration issues, and the diminishing window within which each process is expected to perform, the best one can hope for is that the rate of price increases due to development costs would decelerate.
In addition to the benefits accruing to the slurry and pad suppliers, as well as the industry in general, the facility would be a resource to the university research programs and other industry-funded research at Albany for process integration and feasibility demonstration.
Are companies willing to pony up the cost of one machine to participate? Does the industry really want a greater number of consumable suppliers from which to choose? Support for this facility lowers the entrance barrier for smaller guys who struggled to develop a 200mm apps lab, and are still struggling with the recurring nightmares of a 300mm lab. C3PO can give the industry those choices.
Pay-as-you-go service charges can be expected to cover operating and materials costs for C3PO, but financing of the capital is a different story. Government grants that might support this sort of capital proposal are rare. The way to make C3PO happen is for the fab customers to recognize that this concept is in the industry's best interest, and to modify this proposal to get the kinks out. Some portion of International Sematech (ISMT) dues could perhaps be allocated to support the initial capitalization, and a membership fee for non-ISMT fabs could be assigned.
A nose-to-nose discussion with equipment suppliers to make this financial burden as painless as possible should occur. It should also be made clear to consumables suppliers that benchmarking with de facto standard materials is expected, even across competitive lines. Guidelines would require C3PO users to share data with those fabs and equipment suppliers who subsidized the facility.
This proposal is being discussed with Albany NanoTech. Alain Kaloyeros, dean of Albany's new School of NanoSciences and NanoEngineering, and executive director of Albany NanoTech, said "This is a tremendous opportunity to elevate industry-university-government collaboration to a fully integrated development and deployment approach. Such a facility would make the latest CMP materials and equipment accessible to advanced university research, while at the same time providing a cost-containment mechanism for commercial product development that the industry sorely needs for next-generation semiconductors."
It is my intention — with EKC's blessing — to serve as champion for moving this proposal forward with Albany NanoTech, ISMT, its member companies, equipment suppliers, and consumables suppliers, until another organization is committed to carry the torch.
Michael A. Fury
EKC Technology Inc.
Michael A. Fury is VP of R&D and engineering at EKC Technology Inc., PO Box 3703, Hayward, CA 94540-3703; ph 510/780-1340, fax 925/973-1002, e-mail [email protected].