Sustaining the Leading Edge: Evolving UHP Gas Systems for Advanced Nodes
ABSTRACT
As semiconductor manufacturing advances into the sub-2nm regime with the adoption of Gate-All-Around (GAAFET) architectures, process complexity has grown significantly. The transition from planar and FinFET geometries to multi-bridge channel devices introduces a 20% to 30% increase in front-end-of-line (FEOL) processing steps, driving fab facility capital expenditures to 8–12% of total greenfield investments exceeding $10B. Because ultra-high purity (UHP) bulk and specialty gases make direct surface contact with wafers across nearly all processing steps, gas delivery systems are facing new challenges regarding design and operations.
This presentation outlines some of the shifts required in facilities UHP gas systems to support leading-edge nodes, addressing four interdependent vectors: purity strict requirements, surging per-wafer gas consumption, critical particle filtering, and industry-wide sustainability goals.
Overcoming these bottlenecks requires co-design across chip manufacturers, gas suppliers, and equipment vendors. Narrowing statistical process control (SPC) bands without driving unsustainable CapEx demands targeted purification. Furthermore, stringent Management of Change (MOC) and Process Change Notification (PCN) governance throughout raw material sub-suppliers is critical. This work covers some of the challenges in supporting an evolving fab infrastructure to sustain Moore’s Law economically and reliably.
BIOGRAPHY
Gustavo Albuquerque is a Senior Gas Systems Engineer in the technology development group at Intel Corporation, where he built an extensive expertise in semiconductor manufacturing gas distribution infrastructure. Over his career at Intel, Gustavo has several contributions in designing, building, optimizing, and operating robust gas systems to support advanced technology node scaling. He holds a Ph.D. in Chemical Engineering from Oregon State University and is passioned about process safety and the continuous innovation implemented to support Moore's Law challenges.