This is a Request for Information (RFI) by the Defense Advanced Research Projects Agency (DARPA) regarding the development of capital equipment to address the needs of Low-Volume Nano and MicroElectroMechanical Systems (N/MEMS) fabrication and manufacturing. The fabrication of N/MEMS devices has historically been accomplished using equipment originally designed for the semiconductor (e.g. CMOS) industry. The availability of this equipment has greatly assisted in the development of N/MEMS technology. However, in addition to some differences in performance requirements, there is a fundamental difference in terms of production volume between the needs of military N/MEMS products and the capabilities of commercial semiconductor equipment. Military systems typically require from 10-100,000 devices per year. Semiconductor tools using four-inch wafers are designed to produce 1,000,000 units per year or more. Thus, even using out-of-date four-inch equipment, N/MEMS production facilities are oversized for military devices by 10x to 100,000x. The semiconductor industry has moved on to six-inch, eight-inch and 12 inch wafers. As this happens, and as four-inch equipment becomes increasingly hard to obtain, the gap between the needs of military production and the capabilities of available equipment will only continue to increase. In addition to capital costs, the unwarranted scale of semiconductor capital equipment results in substantial additional maintenance costs, drives up materials costs, requires additional staff to maintain, and requires additional rent and utilities. Even if one obtains capital equipment for free, the maintenance costs (which are often about the same as depreciation costs) can make owning such equipment unaffordable. The MEMS industry has worked to address this problem by moving to a "MEMS Foundry" model. In this model, equipment is owned by one company who performs fabrication services for a range of other companies. This approach can help distribute costs to some extent, but also creates additional issues. First, N/MEMS devices, unlike semiconductor devices, each require their own process. These processes are often incompatible (e.g. use of gold or other matererials), meaning that it is sometimes impossible for multiple devices to be run at a single facility. Second, for devices that are compatible, production time ends up being dominated by changeover time. That is, while it may only take 1% or less of a facility's capacity to produce a year?s supply of product, it may take weeks to change the facility over from one process to another. Thus, one's ability to spread costs over multiple devices is dominated by changeover costs, not production costs. These costs only get worse as we move to larger wafers sizes. Third, because each product is such a small percentage of the overall production mix going through a facility, and since that product mix is always changing, it becomes very difficult to maintain uniform process conditions and repeatable production. This lack of uniformity impacts device performance, device reliability, and can create significant supply interruptions. In short, though N/MEMS has benefited from the existence of semiconductor capital equipment, continuing to rely upon semiconductor equipment to produce N/MEMS device, particularly military N/MEMS devices, may be untenable. DARPA can envision two potential solutions to this dilemma. A partial solution would be to create equipment for use by MEMS foundries whose performance is inherently uniform and repeatable when exposed to a continuously varying product mix. While this approach does not address the issue of process compatibility, it would enable compatible processes to be run with higher performance and reliability within a single facility. Such an approach may require new strategies for process control, monitoring, and simulation. DARPA seeks revolutionary concepts for processing tools to enable a new paradigm for high process mix manufacturing for N/MEMS components. DARPA requests individuals with ideas for creating such inherently flexible, affordable, and repeatable N/MEMS capital equipment submit a brief description of their ideas by email, by the published response date, to Dr. John D. Evans at John.Evans@DARPA.mil. A second and more comprehensive approach is to develop a set of capital equipment that is sufficiently inexpensive that a complete tool set can be devoted to each process. With such low cost capital equipment, each tool set would see only once process, and thus process compatibility is not an issue. Additionally, because each tool is only required to perform a single process, process repeatability can be dramatically improved. While the process performance for such tools would have to be comparable to existing equipment, throughput could be dramatically reduced. This RFI seeks to solicit ideas relevant to the creation of a tool set specifically designed for the production of low-volume N/MEMS devices. This equipment would be designed to: 1. Decrease the cost of capital equipment by 10x relative to existing new four-inch equipment. 2. Target a production throughput of 10,000 units per year (square centimeters per year). This is approximately 100x lower than current four-inch facilities. Where a process requires a piece of equipment be used multiple times during the execution of a process, the tool must be appropriately sized to accommodate such use. 3. Maintain process performance relative to existing N/MEMS equipment (e.g. 1-2 um linewidths for contact lithography). 4. Maintain "hot-lot" turnaround times of less than 4 weeks (layout to devices). Hot lot turnaround time is defined as the time it takes to execute a lot of wafers (e.g. 5 wafers) where the lot never has to wait for equipment to become available. Additionally, this equipment would be designed to meet the following specifications: 1. Facilities. Tools will be completely self-contained and require no external facilities other than HEPA filter coverage, house ventilation, vacuum (for wands, etc.), nitrogen, Clean Dry Air (CDA), De-Ionized (DI) water, and general DI water waste (presumably to carboy). 2. Footprint. Tools will fit within the footprint of one (or more) standard pallets (e.g. 48 x 40 inches). 3. Transportability. Tools will be transportable using a standard forklift, so that they can be easily relocated or transported at end of a each project. 4. Set-up / Take-down. Tools shall be able to be packed up for shipping in under 4 hours, and be able to be unpacked and put into service in 4 hours. This would enable a complete production line to be assembled in approximately one week. 5. Access. Tools will require only front access for normal operation. Tool will roll fore-and-aft under human power for maintenance. These requirements enable a significant decrease in total cleanroom footprint for each tool 6. Power. Tools will operate on standard 110V or 220V power, and come with standard plug. 7. Consumables. Dirty consumables (e.g. gas bottles) will be stored within the footprint of the tool, preferably near the floor. Consumable storage will be appropriately sized to the throughput of the tool. 8. Chemical waste. Chemical waste will be sent to carboy (or treated) and contained within the footprint of the tool, preferably near the floor. Chemical waste storage will be appropriately sized to the throughput of the tool. 9. Safety. Tool will include all necessary safety systems, such as gas monitoring systems. To this end, DARPA is soliciting strategies for realizing this equipment. Though DARPA is interested in all classes of capital equipment, the following ten areas are of particular interest. 1. Deposition (e.g. PECVD / CVD) for silicon dioxide, silicon nitride, and polysilicon. 2. Deposition (e.g. sputtering) of metals. 3. Reactive Ion Etching (RIE) of silicon dioxide, silicon nitride, polysilicon, and metals. 4. Deep Reactive Ion Etching (DRIE) of silicon and xenon difluoride isotropic silicon etching. 5. Scanning Electron Microscopy (SEM) & Critical Dimension (CD) measurement. 6. Film Thickness Measurement 7. Patterning (e.g. lithography). Critical dimensions of 1-2 um. 8. Wet processing. (e.g. spin coating, bake, develop, wet etch, wafer cleaning). 9. Wafer bonding. 10. Facilities. An integrated package of equipment providing vacuum (for wands, etc.), house nitrogen, clean dry air (CDA), de-ionized (DI) water, and DI water waste (presumably to carboy). DARPA solicits potential performers to submit ideas for creating such low-cost capital equipment for use by the government in formulating a potential formal solicitation. DARPA requests that those interested in this topic indicate their interest immediately by email to Dr. John D. Evans at John.Evans@DARPA.mil. DARPA further request that interested parties provide, by the published response date, a submission describing a strategy for realizing a 10x decrease in cost for a specific piece of equipment. Performers should provide a basis for their cost analysis (such as a hypothetical Bill of Materials), and a comparison of that cost with the cost of the existing equipment that their new equipment would replace. Performers should also identify the technical (i.e. research) challenges inherent in realizing this capital equipment. The Government does not intend to award a contract on the basis of this RFI or to otherwise pay for the information solicited. Responses should be limited to five pages for any given concept. Any proprietary concepts or information should be clearly identified as such. Input on technical aspects of the responses may be solicited by DARPA from non-Government consultants/experts who are bound by appropriate non-disclosure requirements. For all RFI responses, an additional, non-proprietary cover page is also requested identifying your company name, technical point of contact and contact information. RFI responses should be received no later than 4:00pm (Eastern) October 30, 2006. RFI Responses should be submitted in a commonly used electronic format and may be e-mailed or mailed on a CD-ROM to: Dr. John Evans, Program Manager DARPA/MTO 3701 N Fairfax Dr. Arlington, VA 22203 John.Evans@darpa.mil
