ISMI Outlines 450 mm Wafer, NGF Roadmaps
ISMI managers described progress at the 450 mm wafer Interoperability Test Bed, and described the Phase 2 roadmap at last week's ISMI Symposium on Manufacturing Effectiveness. Also, the Next Generation Factory program at ISMI is continuing work on cycle time improvements for existing and greenfield 300 mm wafer fabs, including support for 12-wafer lots.
David Lammers, News Editor -- Semiconductor International, 10/27/2008
|
Despite the downturn, the 450 mm wafer program is planning a process demonstration line, while the Next Generation Factory (NGF) program is targeting cycle time improvements for 300 mm fabs, managers said at last week’s ISMI Symposium on Manufacturing Effectiveness, held in Austin, Texas.
|
| Tom Jefferson, 450 mm program manager, ISMI |
The 450 mm wafer processing demonstration line planned for 2010 will be aimed at 32 nm half-pitch technology, with the 2012 pilot lines at several major device makers targeting 22 nm half-pitch patterns, said Tom Jefferson, 450 mm program manager at the International Sematech Manufacturing Initiative (ISMI, Austin).
Jefferson said technology targets for the demonstration and pilot lines came after caucusing the members, and represent the most important recent decisions made at ISMI. “The equipment vendors have been asking us what node they should target their R&D efforts toward,” he said.
ISMI, which includes 17 device makers among its members, will build a demonstration line by 2010, combining wafer handling equipment with the first 450 mm processing tools. That demonstration line will have 32 nm DRAM stagger-contacted Metal 1 half-pitch technology as the target. Then, plans call for the “IST” companies — Intel Corp. (Santa Clara, Calif.), Samsung Electronics Co. (Seoul, South Korea), and Taiwan Semiconductor Manufacturing Co. Ltd. (TSMC, Hsinchu, Taiwan) — each to build 22 nm pilot lines at their respective facilities in 2012. The demonstration and pilot lines would not aim to create finished devices, said Jefferson, an Intel assignee to ISMI.
|
| ISMI seeks to develop a 450 mm wafer demonstration line by 2010. |
On the 300 mm side of ISMI’s program, Olaf Rothe, NGF program manager, said the NGF program has focused on using adaptive scheduling, small lot sizes, and single-wafer processing to reduce the cycle time in fabs. Although several of the NGF productivity enhancement techniques could be introduced into today’s 300 mm fabs as upgrades, the long-term target is to incorporate the NGF recommendations into greenfield 300 mm fabs with new automated material handling systems, he said. “We are aiming at a 50% cycle time improvement and a 30% cost improvement in NGF fabs,” said Rothe, an Advanced Micro Devices (AMD, Sunnyvale, Calif.) assignee to ISMI.
The NGF program seeks to develop the ability to apply a different process recipe to each wafer in a 25-wafer lot, a goal that was attempted by IBM and other companies a decade ago but that was largely abandoned because of complexity issues. By 2010, the NGF program seeks to develop detailed standards for processing 12-wafer lots, Rothe said. A new NGF roadmap may be published around the time of the SEMICON Japan show in December, he added.
|
| Scott Kramer, vice president for manufacturing, Sematech |
The NGF program has been spearheaded by AMD, with support from a list of equipment vendors and other adherants that prefer to concentrate on improving the productivity of 300 mm fabs rather than crossing over into the 450 mm wafer size. The ISMI program’s position is that the NGF goals of improving cycle time, predictive preventive maintenance (PPM), and other productivity enhancement techniques can first be applied to 300 mm wafers and then applied to 450 mm wafer fabs, said Scott Kramer, the Sematech vice president of manufacturing who heads up the ISMI program.
The ISMI members have funded the Factory Integration Interoperability Test Bed (ITB), aimed at developing standards for wafer handling, wafer carrier and metrology vendors. The ITB occupies 2500 ft2 in the Silicon Valley Technology Corp.’s facility in Austin, the former Advanced Technology Development Facility (ATDF). About 1500 ft2 of the 2500 ft2 is occupied now.
Jefferson said the seven companies participating in the ITB include Asyst Technologies Inc. (Fremont, Calif.), Brooks Automation Inc. (Chelmsford, Mass.), Entegris Inc. (Chaska, Minn.), Fixeon LLC (Austin), Genmark Automation Inc. (Milpitas, Calif.), Gudeng Precision Industrial Co. (Taipei, Taiwan) and H-Square Corp. (Santa Clara, Calif.).
Jefferson said the automation vendors have created prototype front-opening unified pod (FOUP) carriers for the 450 mm test wafers, and have determined that the 10 mm pitch (the distance between the wafers) within the carriers can be maintained, simplifying the transition from the 300 mm carriers and robots. Although the 450 mm wafers have a 2.25× larger area than 300 mm wafers, they are expected to be only slightly thicker, with a 925 µm wafer thickness believed to be achievable. A fully loaded 25-wafer carrier is expected to weigh 45 lb.
“The wafer pitch is critical for the automation vendors,” Jefferson said. “After our tests, we believe the 10 mm pitch is very achievable despite the fact that the wafer is a little thicker. We have tested multiple carriers and multiple wafers, and we believe the 10 mm pitch is still achievable.”
Also, the ITB has created a test wafer bank that has loaned out ~120 polycrystalline test wafers to vendors for their 450 mm R&D programs. The ITB got its first test wafer in March, and the rising level of test wafer loans is an indication of increasing R&D interest by the equipment vendors, according to Jefferson.
One project involved determining the level of sag in the polycrystalline test wafers, and then determining if sag in the single-crystal wafers would be the same. The test involved cutting out a 300 mm portion of the 450 mm sintered wafers and then comparing it to an existing 300 mm single-crystal wafer, an experiment that Jefferson said indicates that sag will be nearly identical for the polycrystalline and single-crystal 450 mm wafers.
Next year, ISMI’s test bed plans to create a wafer bank of single-crystal silicon wafers, and work with metrology vendors to develop inspection tools for particle detection, geometry tools and others. Later next year, the test bed expects to receive early blanket thin-film deposition equipment, Jefferson said, and then move on to a “basic oxide patterning cell” in 2010.
The early processing tools could have “relaxed” processing capabilities during the test wafer phase, according to Jefferson. At the ISMI symposium, he published a list of the priority tools sought for the early stages of the test bed as oxide vertical furnace, PECVD dielectric, dielectric dry etch, dielectric CMP, PVD metal, LPCVD nitride, wet cleans, dry strip asher, particle detection metrology, film thickness metrology, and CD measurement.
For the alpha/beta demonstration lines planned for 2010 at ISMI, the priority tools were listed as 193 nm dry scanner, lithography track, PECVD dielectric, dielectric CMP, dielectric dry etch, electrochemical copper plating, high-energy ion implant, LPCVD silicon nitride, wet cleans, and bare wafer particle detection.
Asked if the economic downturn has dampened support for the 450 mm wafer transition, Kramer said, “This is a long-term program. Momentum has been building, but there are still many open issues. The members are supporting the proposals for what we plan to do at the test bed, so the mood is positive now — more positive than six to nine months ago. We are always updating our plans, continuously evaluating things. If the current financial problems affect companies’ plans, we will adjust accordingly.”
|
| Joe Draina, director, ISMI |
ISMI Director Joe Draina said, “The team at ISMI has met with many of the equipment suppliers. Many of the suppliers are not saying no. They are not saying they are there yet, but they are not saying ‘Forget this.’ They are saying, ‘We need to see some orders so that we know who is committed.’”
-
Changing to 450 mm Wafer substrate is really quite expensive now, with all the various processes used. Have they thought about SOI?
doubtful - 10/28/2008 10:08:00 AM CDT -
"Where are the OEM's?"
"Where are the OEM's?" - 10/27/2008 3:01:00 PM CDT -
It is hard to imagine any equipment supplier that would consider 450mm research and development aimed at 32nm when they are already far along on developing that process on 300mm. It is unlikely that there would be any 32nm or 22nm orders placed on 450mm. 16nm seems a more likely target, if the cost/benefit proves out.
Really? - 10/27/2008 2:38:00 PM CDT
ISMI Readies 450 mm Test Bed, Wafer Bank
03/13/2010ISMI to Report Progress at SEMICON West
03/13/2010ISMI Readies 450 mm Equipment Metrics
03/13/2010ISMI to Report 450 mm Progress at West
03/13/2010ISMI 450 mm Program Moving to Next Stage
03/13/2010
