Erik Smith, President, Qcept Technologies, Atlanta, GA USA

The semiconductor industry will be challenged in 2009. The economic crisis in the U.S. and around the globe has led to weakened demand for PCs and consumer electronic products, as well as corporate IT investments, all of which fuel IC consumption. Amid the gloomy outlook on IC demand in the coming quarters, many chip manufacturers are cutting back on their investments to build capacity. However, it is at this time — when demand is at its softest — that the wisest of chip manufacturers are continuing to make strategic investments in new manufacturing technologies and methodologies in order to position themselves to quickly ramp up on new chip designs when semiconductor demand inevitably swings upward once again.

In past downturns, the nature of these technology investments skewed heavily toward design shrinks as the primary driver for device performance improvements. Now however, we are seeing the escalating cost of lithography pushing the leading chip manufacturers toward the integration of new materials into their manufacturing processes as an important lever. These new materials, such as high-k and low-k dielectrics, metal gates and advanced photoresists, improve the performance of their devices and increase average selling prices (ASPs) on their shipped products. As always, each new material and the added process steps associated with its introduction requires yield learning in order to bring the new process into maturity. Thus, yield management remains an investment priority for the top semiconductor manufacturers regardless of the stage of the cycle in which the industry finds itself.

However, in the same way that there is a rethinking of how to drive performance improvements, a rethinking of traditional yield management strategies is also underway. These new materials and processes rely on very tight tolerances on surface condition in order to reap their performance benefits. This increases the importance of controlling non visual defects, such as submonolayer organics and metallics, and process-induced charge, which arise from repeated surface prep and cleaning process steps, and which cannot be detected through traditional inspection methods.

Just as we are seeing “out of the box” thinking in terms of how to drive device performance, we are seeing the same in terms of yield management. By making technology investments now, the industry will have the infrastructure required to raise volume yields in this new era of materials and process innovation that will fuel performance growth in 2009 and beyond.