News

Japan's CASMAT may be the model for future R&D investment.

Semiconductor packaging materials are in a dynamic phase. Organics rule memory devices and CPUs for PCs. Many devices are migrating to wafer-level packages. What are the drivers for this change? Who will develop these new materials? Who will underwrite the R&D cost?

Semiconductor material developments are driven by performance needs. The introduction of low-k dielectric materials is an excellent example. These materials will see continued improvements? Successful material introductions cannot take place without the coordination of back-end assembly and test changes. With the first-generation low-k materials, the conventional test and assembly process created stresses that resulted in delamination between the low-k material and the silicon. Issues included concerns with damage from probe cards as well as stresses caused by the materials used in the assembly process. For wire-bond parts this meant changes in mold compounds, and flip chips required new underfill formulations. As future generations of low-k materials are rolled out, material changes in the assembly process must also be considered. While companies like LSI Logic have worked closely with assembly operations and material and equipment suppliers, this is an expensive process. Companies such as Intel and IBM still have the financial and manpower resources to tackle these issues internally, but many fabless companies do not have extensive internal knowledge to deal with the required changes. Subcontract assembly service providers do not have the profit margins to support development of new semiconductor packaging materials. Thus, semiconductor foundries must play a greater role not only in IC design, but also in packaging, assembly and even materials development.

Beware the Consumer

While the PC segment remains important in terms of the dollar value of semiconductor content, consumer products - mobile phones and other handheld communication devices - drive unit volumes. Packaging for consumer electronics is driven by continued demand for smaller and portable products with greater functionality. At the same time, these applications are also driven by performance requirements of increased data rates for image processing and gaming. The changes in the semiconductor packaging requirements impact test, while assembly and materials development necessitates advances in wafer thinning, die attach materials and chip bonding.

Historically, many semiconductor packaging material developments took place in large, vertically integrated companies such as IBM and AT&T. In addition, many energy companies developed materials for electronics. These days, the vertically integrated companies are changing their structures and the energy companies have divested or closed their electronics materials divisions. Many major suppliers today are in Japan. While many Japanese companies still maintain large R&D budgets for materials, even these firms are concerned over the resources required to develop next-generation material sets. Joint research activities are one way to help share the expense and burden of new material developments. At the SEMI Strategic Materials Conference held in Half Moon Bay, CA, in January, the Japanese consortium CASMAT was described. Not only does this research organization, funded by Japan's government, explore materials for semiconductor fabrication processes, but it also researches materials for packaging and back-end assembly that would affect the introduction of technologies in the front-end.

Packaging and assembly has become the gating factor to the performance potential of new semiconductor designs. The semiconductor materials market for IC packages has grown over the years and is of significant size (Table 1). The market is expected to grow from $12 billion in 2005 to $19.5 billion by 2010 based on projections published in the new Global Semiconductor Packaging Materials Outlook. But without the continued introduction of materials and coordinated assembly process development, the industry will face limits. Limited investments in production capacity for laminate substrates and even materials to produce substrates such as BT resin have caused shortages and are impacting the entire industry. The impact of underinvestment in the development of new materials will be even greater.

E. Jan Vardaman is president of TechSearch International, Austin, TX; jan@TechSearchInc.com. Her column appears semimonthly.