The semiconductor memory industry, driven by the explosive growth of Generative AI and High Bandwidth Memory (HBM), demands pattern fidelity at scales that continually redefine the limits of lithography. Over the past four decades, the industry has pursued finer patterning through ever-shorter wavelengths — from i-line through KrF and ArF to Extreme Ultraviolet (EUV) — and the transition to High-NA EUV is now underway, with SK hynix among the first to prepare for high-volume manufacturing: the company has already progressed through tool installation, process setup, and initial pattern verification. Yet this transition is far more than a tool upgrade. Increasing the Numerical Aperture (NA) delivers superior resolution but sharply reduces Depth of Focus (DOF), and unlike previous node transitions, nearly every process element — resist, mask, optics, system — demands comprehensive re-validation. High-NA EUV is not simply a technology for printing smaller patterns; it is a technology that must overcome reduced DOF to achieve higher resolution.

A. The DOF-Selectivity Dilemma: Why Thin CAR Resist Alone Is Not Enough

Among the representative strategies for recovering DOF margin in High-NA — Source, Mask Optimization (SMO), Focus Drilling, and PR thickness reduction — thinning the photoresist (PR) is the most intuitive and practical, and indeed a long-standing industry trend driven by pitch shrinkage and pattern collapse concerns. However, it carries a critical side effect: after patterning, the remaining PR is too thin to protect the underlying layer during etch. Metal Organic Resist (MOR), with its inherent etch resistance, offers a natural escape — yet in many critical layers, MOR cannot yet fully replace chemically amplified resist (CAR). Simply requesting a "thinner, stronger material" from resist suppliers has clear limits. This is the question that motivated a new approach: how can thin CAR PR be made robust against etch while still transferring patterns cleanly to the layer below? The answer is not a new material, but a new post-process.

B. PR Freezing: Restructuring Patterned Resist into a Carbon Hardmask

SK hynix's proposed solution "PR Freezing" applies argon implantation to patterned EUV PR, reorganizing its internal bonding structure in situ. The concept builds on prior observations that implantation reduces defective hydrogen bonding within resist films, and the experimental results confirm the mechanism at scale. X-ray Reflectivity (XRR) analysis shows a clear density increase after freezing, directly correlated with etch rate improvement across various types of CAR PR. Fourier Transform Infrared Spectroscopy (FT-IR) reveals the chemistry behind it: the organic bonds that served essential roles during patterning — -C=O, -C-H, -C-O — are substantially depleted after implantation. As an inert species, Ar breaks the existing C, O, and H bonding networks, and recombination preferentially favors C-C bonds (348 kJ/mol) over C-H (410–439 kJ/mol) and C-O (351–380 kJ/mol). The resulting graphite-like, carbon-enriched film effectively converts patterned resist into a functional amorphous carbon hardmask — restoring etch selectivity even when the remaining PR is too thin to survive pattern transfer.

C. Defining the Freezing Window: Balancing Loss Against Gain

PR Freezing is not without cost: the process consumes a measurable portion of the PR film upfront, with thickness loss growing at higher implantation doses. Moreover, the etch benefit saturates — beyond a certain dose, further increases yield no additional improvement, with the saturation point varying by resist composition. Effective implementation therefore requires reading two parameters together: how much PR is lost initially, and how much the etch rate improves — only when both are balanced does an effective freezing zone emerge. Early data from the PR Freezing process show that, despite the expected post-freeze thickness reduction, the DI-FI bias — a direct indicator of etch rate improvement — decreases without any degradation in CD uniformity. By mapping how initial material properties and process variables shape final film characteristics and patterning performance, a robust operating range can be established even at reduced PR thickness.

The convergence of structural reorganization physics, carbon-like film transformation, and a precisely bounded process window positions PR Freezing as a key enabler for High-NA EUV patterning. By decoupling DOF margin from etch selectivity, it allows us to pursue aggressive PR thinning without sacrificing pattern transfer — a critical step toward stable, high-yield manufacturing in the High-NA era.