July 5, 2007 – Fresh off its recent acquisition of UK-based Epichem Ltd., a provider of high-purity chemicals (e.g. precursors for high-k materials), Sigma-Aldrich is focusing on silicon and compound semiconductor markets, touting a beefed-up division’s product roadmap at SEMICON West. Execs talked with WaferNEWS ahead of the announcements, about the company’s direction, its focus on collaboration, and how to justify value in an increasingly consumer-oriented industry where lower prices mean tighter investment evaluations.
SAFC is Sigma-Aldrich’s custom manufacturing and services division focusing on inorganic materials, accounting for about ~30% of total group’s $1.8 billion in sales — 12% of that ~$550 million is SAFC Hitech, of which two-thirds comes from Epichem. Starting in 2000 SAFC began redirecting efforts to dedicate facilities, sales/marketing, and other processes to different customers, and now the expansion with Epichem is an effort to diversify away from, and provide some balance to, the group’s other core market of pharmaceuticals, noted Frank Wicks, SAFC president. Epichem had previously been the beneficiary of backing from Intel Capital, back in late 2005.
Specifically for the semiconductor industry, SAFC sees penetration in a range of areas, including CVD precursors, barrier layers, high-k dielectrics for advanced gate stacks, low-k dielectrics for IDL applications, and specialty coatings. Geoff Irvine, SAFC’s director of R&D, told WaferNEWS the company sees two insertion points in the silicon industry for the company’s products: collaborative joint development projects for completely new materials, and simply supplying a known chemical to support the supply chain. For example, an improved corrosion inhibitor can be added into an existing CMP slurry set.
SAFC Hitech emphasizes the importance for materials suppliers to be increasingly collaborative with customers, and the need for a much clearer communication interface, noted Irvine. SAFC had a direct access to device manufacturers’ electronics value chains, with a good manufacturing footprint — but lacked customer interfaces with applications development teams, he noted. Those are brought to the table by Epichem, which has deep ties to compound semiconductor manufacturing (e.g., TMI, TMA, and materials with very low oxygen content), but lacked the means to achieve rapid scale-up and commercial manufacturing, Irvine said. Together, SAFC Hitech has “all the pieces put together,” he said: a strong customer interface, application development capabilities to talk with customers and tune products to their needs, and process development and manufacturing scale-up footprints.
Maintaining a collaborative supply relationship from development to manufacturing also helps the supplier make its case for the value of its products/services, a key sticking point particularly in the materials side of the semiconductor equation, Irvine suggested. “Transparency of the supply chain is critical,” he told WaferNEWS. “The end market needs to understand the value of where it’s created.” With the chip industry becoming ever more consumer-based (and seeing ASPs driven down), the need to justify value is increasingly weighted toward early development among suppliers, he said. “The only way to fight that is to address customers directly, what drives their motivations in the marketplace. When you are “supplier 1-5″ operating in the background, you’re not getting very much information back.”
For example, a devicemaker may want a molecule that integrates well in a gate stack, and Solution A is its top choice offering the best desired characteristics, but it’s not economically viable to manufacture for the supplier. A chemicals supplier with a collaborative model, however, might be able to suggest going in the direction of Solution B instead, where results may be close but not quite as good, but margins will be more attractive on a materials basis, and more easily scaleable in manufacturing.
Irvine noted that SAFC Hitech has picked up new business from the Epichem deal, e.g., high-brightness LEDs, and is now selling into silicon -based logic, ASICs, CPUs, and even solar. Work involves thin-film deposition materials, such as liners for both frontend and backend processing. “We get a lot of inquiries for collaborations, but we have to be careful what we take on board,” vs. what promises can be fulfilled, he noted.
Other major work at SAFC Hitech involves high-k dielectric materials, another area where a collaborative, clear-communicative relationship with customers is key, since it looks like several different high-k material flavors will be required, depending on process characteristics (e.g., different DRAM architectures) and end applications (e.g., high-power or low-power/mobile devices), meaning manufacturing not just one hafnium material, but maybe two or three, Irvine noted. Not to mention increasing challenges with exotic and niche materials and in volume amounts, e.g. for adhesion to aid in diffusion barriers.
One area offering promise to chemicals and materials developers is “combinatory discovery”, a technique pioneered for drug discovery that promises to speed up the process for finding and characterizing new structures. Early adopters of this method include Dow Chemical, which claims significant benefits in terms of R&D speed and efficiency. SAFC’s Irvine said that although it’s an attractive model for a market such as semiconductors with very long cycle times and extensive capital costs, it’s still difficult to get materials to the proof-of-concept stage. “We’re all still hoping” for the promises of combinatory discovery, but it hasn’t really been embraced, he noted. — J.M.
