Advanced Metrology & Inspection

As semiconductor processes scale down, Metrology & Inspection (MI) plays an increasingly vital role in in-line pattern profile management and defect classification. SK hynix Inc. leads MI innovation through AI and big data, advancing technologies to maximize process reliability and product yield.

MI has evolved in step with DRAM and NAND development. DRAM has sustained performance improvements through cell size reduction; however, as physical limits are approached, process complexity has escalated rapidly with the adoption of EUV lithography and the concurrent development of next-generation 3D DRAM architectures. NAND has transitioned from two-dimensional planar to three-dimensional stacked structures with layer counts now extending into the hundreds, necessitating novel MI methodologies capable of characterizing defects within high-aspect-ratio structures and discerning inter-layer electrical properties.

Aligned with three key trajectories—DRAM scaling, NAND vertical stacking, and wafer bonding—SK hynix Inc. is transforming MI from conventional measurement and inspection toward predictive, high-intelligence paradigms. This vision focuses on building an MI ecosystem that enables smart, autonomous process governance.

Metrology

As memory devices evolve toward higher-density integration and increasingly complex three-dimensional (3D) structures, nanoscale features can become difficult to interpret unambiguously in conventional two-dimensional imaging due to overlapping information along the viewing direction.

Increasing stack heights in 3D DRAM and NAND have exposed limitations in conventional metrology, while phenomena such as wafer warpage present additional challenges. SK hynix Inc. is pushing the limits of optical, SEM, and X-ray technologies, and is building hybrid metrology that integrates high-speed optical measurement, high-resolution SEM analysis, and X-ray internal stress characterization for mutual compensation.

Going forward, metrology must transcend mere measurement to establish systems that fuse and analyze vast multi-modal data in real time through AI, optimizing process variables and preemptively preventing defects.

Inspection

Detection has shifted from surface-defect inspection in 2D planar structures to multi-wavelength optical inspection for identifying internal structural defects and inter-layer micro-defects, driven by 3D high-layer-count stacking and wafer bonding processes.

Electron beam inspection (EBI) provides nanoscale resolution for fine pattern defect detection. To overcome EBI’s throughput limitations, SK hynix Inc. is developing multi-eBeam inspection technology.

Sensitivity enhancement now incorporates AI algorithms for dynamic thresholding and defect classification, complementing hardware advances. Building on these capabilities, SK hynix Inc. is integrating real-time inspection-condition optimization to advance toward fully autonomous inspection process control.