IN Brief:
- Samsung is reported to be targeting its 1nm-class generation around 2030 for High-NA EUV production insertion.
- ASML's High-NA platform raises numerical aperture from 0.33 to 0.55 and provides 8nm resolution.
- Conventional EUV and High-NA processes are expected to coexist as foundries balance resolution, cost, throughput, and process maturity.
Samsung Electronics is reported to be targeting its 1nm-class foundry generation for production use of High-NA extreme-ultraviolet lithography, placing the next-generation exposure technology later in its process roadmap than earlier ambitions around 2nm and 1.4nm.
The reported plan would put High-NA EUV around Samsung’s A10 generation, with 1nm production expected around 2030. Samsung has already installed High-NA development equipment at its Hwaseong campus, but conventional EUV is expected to carry more of the manufacturing workload through preceding nodes while supporting masks, resists, metrology and process integration continue to mature.
ASML’s High-NA EXE platform increases numerical aperture from the 0.33 used by current NXE production systems to 0.55. The EXE:5000 provides 8nm resolution and can print features 1.7 times smaller in a single exposure than NXE systems, reducing the need for multiple patterning on some advanced structures.
Single-exposure patterning can remove process steps that would otherwise require repeated lithography and etching. Fewer steps can reduce cycle time and some sources of variation, but the scanner is only part of the manufacturing equation. High-NA also requires changes elsewhere in the process flow, including masks, pellicles, photoresists, computational lithography and inspection.
The EXE optical system differs physically from conventional EUV as well. Larger anamorphic projection optics provide the 0.55 numerical aperture while retaining the industry’s established reticle format, but they halve the exposure field compared with NXE systems. ASML compensates with faster wafer and reticle stages to maintain productivity.
Those changes mean a working development scanner does not automatically justify insertion into volume production. Foundries have to consider overlay, defectivity, throughput, yield, maintenance and cost per wafer alongside nominal resolution. A conventional EUV process using multiple patterning can remain preferable if its higher process complexity is outweighed by mature equipment and better economics.
The leading foundries are consequently approaching High-NA on different schedules. Intel has moved furthest towards production use, while TSMC has indicated that several planned nodes can continue without depending immediately on High-NA. Samsung’s reported 1nm timing would extend its use of conventional EUV further while reserving the higher numerical aperture for structures where the additional resolution becomes more difficult to avoid.
Development spending does not stop while production insertion is deferred. High-NA creates new requirements across resists, masks, pellicles, inspection, metrology and computational design, and those technologies need qualification before a foundry can build a stable process around the scanner. Equipment suppliers therefore remain involved well before the first volume wafers run through an EXE production layer.
The decision also illustrates why nominal node labels reveal less about the underlying manufacturing flow than they once did. A single process generation can use different lithography methods for different layers, selecting conventional EUV, High-NA EUV or other patterning techniques according to geometry, yield and cost. High-NA adoption is therefore likely to be layer-specific rather than an immediate replacement of 0.33-NA EUV across an entire process.
Samsung has not issued a formal public production announcement confirming the reported 1nm insertion point, so the roadmap remains subject to change as development continues. Current reporting nevertheless points to a cautious introduction around 2030, with conventional and High-NA EUV operating alongside one another rather than one technology disappearing as soon as the other reaches the fab.
