First Workshop for Solder Joint Design

NSF Geometry Center

May 31 - June 2, 1994

Workshop Report


From:	Carol Handwerker, NIST 
	carolh@enh.nist.gov 
	(301) 975-6158 

The dominant product and technological drivers in almost all commercial electronics systems are cost, performance, and reliability. While these same drivers are important in determining what electrical and optical interconnections to use in a particular design, industry has been hampered from optimizing interconnection choices because there are no tools to integrate design/performance with manufacturing and reliability. As dimensions of microelectronic circuits become smaller and smaller, the importance of such design tools will continue to increase.

Two examples that indicate the need for such tools for solder systems are: fine pitch solder joints and alignment of optical interconnections. Defects in solder joints are increasing in number and severity as the pitch becomes smaller. The choices of component lead and board geometries and assembly processes are made without being able to evaluate the probability of forming short or open circuits during assembly or to predict the reliability of the resulting fine pitch solder joints. Solder is also being used for alignment of optical devices without a quantitative understanding of the degree of alignment possible in the design. Furthermore, the stability of alignment is dependent on the creep and fatigue of the solder alignment system. Inspection and repair of individual solder defects are expensive, labor-intensive operations that can no longer be performed in the current economic climate and for small size scales in electrical and optical interconnect systems.

The lack of integrated design, manufacturing, and reliability modeling was identified as a potential show stopper by industry in the Semiconductor Industry Association (SIA) Semiconductor Technology Roadmap. Development of a wide range of system models for packaging is required to fulfill this need, and extensive and reliable manufacturing data and standards are needed in order to allow validation of these system models. While some design and reliability models are used by industry, there are no manufacturing models in use.

Over the past five years, several industrial, government, and academic groups have independently recognized the need for an integrated design system for solder joints capable of analyzing specific geometries in terms of manufacturability and reliability. Each of these groups has a major research program in solder joint design for performance, manufacturing, and reliability. These groups and principal investigators include: Digital Equipment Corporation: Gary Friedman, National Institute of Standards and Technology (NIST): Carol Handwerker; Marquette University: Steve Heinrich, University of Colorado - Boulder: Y.C. Lee, State University of New York - Binghamton: Tim Singler; Massachusetts Institute of Technology: Julian Szekely, University of Technology - Loughborough: David Whalley and Sandia National Laboratories: Fred Yost.

Representatives from these research groups were invited to meet for a Workshop on Solder Joint Design at the NSF Geometry Center, Minneapolis, MN, from May 31 to June 2, 1994, sponsored by the NIST Center for Theoretical and Computational Materials Science. The purpose of the Workshop was to discuss the modeling approaches that each has taken and to examine whether these organizations could join together to create an integrated design tool for design, manufacturing and reliability for solder joints.

The meeting began on Tuesday afternoon (May 31st) with Ken Brakke's introduction to the Surface Evolver program and a brief orientation to the Geometry Center. The participants spent the rest of the afternoon transferring their simulations to the Geometry Center computers. On June 1 morning each research group described its approaches to solder joint design and the types of modeling/simulations they do. For the remainder of that day and the following morning, the participants demonstrated the range of their simulations on the computers at the Geometry Center. Some researchers using different computational approaches ran their simulations with identical geometries to get a quantitative comparison of their models' predictions.

The workshop ended with a general affirmation that the participants want to work together as a group to generate design tools for solder in electrical and optical interconnect systems. In order to publicize the existence of this Working Group on Solder Joint Design (name may change) in the electronics community and to solicit collaboration with industry, the following steps will be taken:

  1. Produce a workshop report that can be distributed as an information/publicity document for the group. This report will include a description of each group's work, the viewgraphs used in the presentations, and the work plan of the group.
  2. Provide an overview of the group's activities on MOSAIC, including movies of simulations as well as still shots of the final results.
  3. Announce to the electronics industry the existence of the group through regular mail and electronic mail and tell them how to gain access to the hard copy report and to the MOSAIC files.
  4. Present the Working Group's plans and the research results of each individual organization at a session on solder joint design for manufacturing and reliability at an electronics association meeting. This will require that the group organize the session. An output of the group will be a set of archival journal papers to be published together. This will complement the first information document.
  5. Hold another workshop at the Geometry Center to which other industrial partners, such as Ford, Hughes, Motorola, Texas Instruments, and Aerospace Corporation, will be invited.
We are grateful to the NSF Geometry Center for its support of this program. This workshop would not have been possible without having access to the personnel and facilities of the Geometry Center. The work of Ken Brakke is central to the program. His presence at the workshop helped make it a success and his willingness to continue collaborating with the group will be essential to fulfilling the group's long-term goals. The collegial atmosphere of place, with its helpful staff members, such as Angie Vale, and interesting visiting mathematicians, such as Rob Kusner and Jean Taylor, makes it a place where stimulating discussions occur naturally, and where work can get done. NIST's Center for Theoretical and Comptational Materials Science has been formed using the NSF Geometry Center as a model of how to get people to really work together on common modeling problems.

We will return to the Geometry Center for a week-long workshop in 1995. Some travel funding will be available from NIST and the Geometry Center. This meeting will include some additional modelers from industry in addition to those at the first workshop.

If there is any funding available from the Geometry Center for travel or lodging and meals, please consider us for such funding. The NIST Center for Theoretical and Computational Materials Science will provide some funding for the workshop, but the exact amount available for the group will not be known until after the beginning of the next fiscal year.

Thank you again for your assistance. I look forward to our continuing cooperation.

Participants