By Lise Laurin, special to SST
The economics of 300mm wafer processing will demand more effective process control of equipment and tools, according to William Rozich, IBM's director of equipment technology. Improved overall equipment efficiency (OEE) will be essential to cover the higher cost of wafers and the tools needed for 300mm production. Rozich points out that the area of a 300mm wafer is 2.25 times larger than that of a 200mm wafer, but the raw material is expected to cost three times as much.
Speaking at an advanced equipment control/advanced process control (AEC/APC) symposium organized by Sematech in Vail, Colorado, in September, keynote speaker Rozich made it clear that IBM wants equipment manufacturers (OEMs) to integrate the necessary sensors on their tools and take responsibility for communications. He stated that companies like Applied Materials, which supply both metrology and wafer processing tools, should be in a good position to do the integration. He gave an example of some cluster tools where, he said, “integration is occurring regardless of the suppliers involved.”
Other integration alternatives, in Rozich's view, are cooperative efforts between several suppliers, and perhaps a merger of suppliers.
The conference covered automated fault detection, run-to-run control, and real-time process control. Most of these control systems require the integration of specialized sensors on the tool. Some of the sensors measure process state (most end-point detectors fall into this category), while others measure actual wafer parameters, such as film thickness. Many of the sensors falling into this latter category are termed “integrate metrology” because they replace existing off-line metrology tools.
IBM's commitment to AEC/APC is demonstrated by its establishment of a 300mm Integration Facility, dedicated to testing and qualifying advanced control systems.
Other attendees, from chipmakers such as AMD, TI, Infineon, and Motorola, including the White Oak Semiconductor facility near Richmond, VA, presented work on both development and production-scale control schemes. Other IC manufacturers, such as Samsung, Lucent, and Intel, also sent representatives, showing the increased interest in advanced control systems.
IBM's Rozich and others tied sensor integration into OEE. In a real-time control project on epitaxial deposition, for example, reduced Cpk, yield improvements, test wafer reduction, and downtime and set-up time reduction resulted in a 5.3% increase in OEE. This work was co-sponsored by On-Line Technologies, Wacker, Applied Materials, and MIT.
Cost of ownership was another, somewhat related, target for advanced control. The implementation of an endpoint detector on an oxide etch process eliminated two process steps, resulting in savings of $10 per wafer, according to Steve Gunther of CETAC. Christopher Bode of AMD presented data on a run to run controller for photolithography overlay that eliminated test wafers and tool qualifications — two major steps in cost reduction.
Applied Materials has taken a unique approach to AEC/APC by using sensors already integrated into its tools. Lalitha Balasubramhanya showed how the pressure-control throttle-valve position (a parameter already available on most vacuum systems) can be used as an endpoint detector for a chamber etch. Terry Reiss, also of Applied Materials, presented an overall equipment index, comprised of existing tool parameters, which has successfully detected numerous faults.
Attendees involved in CMP were quite interested in a paper by Todd Cerni of Particle Measuring Systems describing an on-line, continuous process monitor of slurry particle size. Initial data showed promise that this technique might prevent scratching caused by larger particles in a slurry.
While it seems that all this talk of advances in APC indicates that model-based, real-time control, complete with advanced sensors and integrated metrology must be right around the corner, behind-the-scenes discussions did not paint so rosy a future. While chipmakers such as IBM, TI, and AMD are making tremendous strides in APC integration, they have not been so eager to share the details of their technology with others. In some cases, they are finding that these controls can give them an important edge in manufacturing. This makes it difficult for the OEMs to make much progress on integrating the new technology effectively into tools.
Only a few OEMs say that they are finding a high demand for APC or even sensor integration from fabs, and when there is a desire for these enhancements to tools there is little willingness to pay the true cost for them. An example was an offer to pay $15,000 for sensor technology that the OEM said would actually cost $60,000 or more to add to the tool.
Another stumbling block is in the area of communications standards. While DeviceNet is emerging as a predominant standard, some of the major players (including IBM) are seriously considering an Ethernet-, TCP/IP-based network. In furnace processing, network protocols have been rejected in favor of simple 0-to-5 volt analog signals for feedback control. In spite of the best efforts of Sematech, it appears a true communications standard will continue to elude the community for a few years yet.
Many of the sensor manufacturers are hearing need for their technology, but they aren't getting many orders. They attribute this partly to the recent weakness in the industry. They understand that OEMs don't want to add cost to their equipment without significant demand, particularly in a down market. But another major reason for the slow adoption of APC and sensor integration may be that the engineers making the equipment decisions don't know about the options available to them. Of the 40 papers presented at last year's AEC/APC conference, only a handful made it into print outside of the symposium.
There were some 250 attendees to the little known meeting this year in Colorado, with some of the papers reporting important progress. Organizers say that the conference is set to expand substantially next year.
This story is adapted from WaferNews, Solid State Technology's weekly news briefing for the semiconductor equipment and materials community. For more information, or to request a sample issue, see www.wafernews.com.
Lise Laurin began her career as a process engineer at Intel and has held a number of positions in semiconductor processing and marketing over the last 18 years. She founded Clear Tech in 1996 to provide technical marketing and consulting services to the semiconductor supplier community. She holds a Bachelor of Science degree in Physics from Yale University, and is an active member of the Semiconductor Safety Association and the SEMI New England Committee.