Fractilia, a leading provider of stochastics metrology solutions for the semiconductor industry, has raised concerns over escalating yield losses driven by random patterning variations at advanced manufacturing nodes. In a newly published whitepaper, the company warns that uncontrolled stochastic effects—random molecular-level variabilities in lithographic patterning—are now the most significant barrier to achieving consistent, high-yield production at cutting-edge semiconductor nodes.
According to Fractilia, these stochastic variations are costing manufacturers hundreds of millions of dollars per fabrication plant, with industry-wide losses reaching into the billions. The firm argues that conventional process control methods are no longer effective in managing the unpredictable behavior inherent in advanced lithography processes, particularly those involving extreme ultraviolet (EUV) and high-numerical aperture (high-NA EUV) technologies.
“Stochastic variability is contributing to multibillion-dollar delays in introducing advanced process technology into high-volume manufacturing,” said Chris Mack, Chief Technology Officer at Fractilia. “Yet conventional process control approaches have been ineffective in addressing these random effects. Closing the stochastics gap requires completely different methodologies that device makers need to validate and adopt.”
The whitepaper introduces what Fractilia terms the “Stochastics Gap”—a widening disparity between what can be achieved in research and development environments and what can be reliably replicated at scale in production. While R&D efforts have successfully patterned features as small as 12 nanometers, scaling these innovations into mass production has been hindered by the inherent randomness of stochastic defects, which can impact yield, performance, and reliability.
Unlike traditional process variability, stochastic effects stem from intrinsic randomness in the behavior of light, materials, and atomic interactions used in lithographic processes. These effects have historically had minimal impact on production yields but now account for a substantial portion of the process error budget, particularly as feature sizes shrink and lithography tools become more precise.
Fractilia’s report outlines a roadmap to mitigate these losses, advocating for a combination of precise stochastic measurement technologies, probability-based process control, and design methodologies tailored to stochastic-aware manufacturing. The company also emphasizes the importance of materials advancements and cross-functional industry collaboration.
“The stochastics gap is an industry-wide problem,” Mack added. “This issue can be minimized and controlled, but it all starts with accurate stochastics measurement technology.”
