Papers
Topics
Authors
Recent
Search
2000 character limit reached

Holographic EUV Lithography at 40 nm Resolution

Published 20 May 2026 in physics.optics | (2605.21430v1)

Abstract: Extreme ultraviolet (EUV) lithography is the cornerstone of the fabrication of advanced integrated circuits at the 7-nm node and beyond, but its reliance on multi-element reflective projection optics makes it inaccessible for small-scale research and prototyping. EUV interference lithography (EUV-IL) provides a lensless alternative but is intrinsically restricted to periodic structures. Here we demonstrate EUV holographic lithography (EUV-HL) as a lensless route to arbitrary, non-periodic, curvilinear patterning at the EUV wavelength of 13.5 nm. We introduce an inverse-design framework for computer-generated holograms that captures the dominant physical effects of EUV mask diffraction within a shift-invariant convolution model that is tractable for full mask layouts. Using this framework, we design and fabricate transmissive holographic masks by direct-write electron-beam lithography in hydrogen silsesquioxane, expose them with synchrotron-generated EUV radiation, and print target layouts with critical dimensions down to 40 nm, nearly an order of magnitude finer than the previous state of the art in EUV-HL. The demonstrated combination of sub-50 nm resolution, curvilinear design freedom, and a lensless optical setup establishes EUV-HL as a uniquely flexible tool for nanostructure prototyping at EUV wavelengths, and provides a natural pathway to non-periodic pattern prototyping at beyond-EUV (BEUV) wavelengths, which is currently inaccessible to interference-based methods.

Summary

  • The paper introduces a novel lensless EUV holographic lithography method achieving 40 nm critical dimensions for arbitrary, non-periodic patterns via inverse-design mask optimization.
  • It employs a gradient-based inverse-design framework that models M3D absorber stack scattering and source polychromatism to precisely control the aerial image.
  • Experimental SEM imaging confirms strong pattern fidelity between simulation and the fabricated resist, establishing EUV-HL as a promising tool for sub-50 nm device prototyping.

Holographic Extreme Ultraviolet Lithography at 40 nm Resolution

Context and Motivation

Extreme ultraviolet (EUV) lithography has become crucial for advanced integrated circuit manufacturing, specifically at the 7-nm technology node and below, facilitating continued device scaling. Traditional EUV patterning leverages highly complex, multi-element reflective optics, imposing significant barriers in terms of cost, accessibility, and engineering, especially for non-industrial research and rapid prototyping. EUV interference lithography (EUV-IL), while providing high-resolution, lensless patterning, remains restricted to periodic structures, limiting its utility in device prototyping involving arbitrary, non-periodic geometries.

To address this gap, the paper presents a lensless EUV holographic lithography (EUV-HL) method that enables arbitrary, curvilinear nanoscale patterning at the EUV wavelength of 13.5 nm, achieving critical dimensions (CD) down to 40 nm—demonstrably surpassing prior resolution benchmarks by nearly an order of magnitude (2605.21430).

Technical Approach

Inverse-Design Framework for Mask Optimization

EUV-HL requires the synthesis of designated intensity distributions in the wafer plane via computer-generated hologram (CGH) masks without reliance on projection optics. The authors develop an inverse-design framework that models the mask's physical response as a shift-invariant convolution, capturing three-dimensional (M3D) absorber stack scattering and the effects of source polychromatism. This tractable model enables efficient optimization over full-mask layouts, solving the inverse problem with gradient-based methods. The MSE between the simulated resist pattern and the target layout serves as the optimization loss.

Mask Design and Fabrication

Transmissive masks are fabricated using direct-write electron-beam lithography in hydrogen silsesquioxane (HSQ) on a silicon nitride membrane. HSQ is chosen for its high-resolution patterning capability, EUV irradiation stability, and compatibility as both a resist and absorber, circumventing additional pattern-transfer steps. The mask layout exploits continuous-tone variations, finalized as a single-thickness absorber for manufacturability.

Modeling Critical Physical Effects

Key to accurate aerial image formation are two factors:

  • M3D effects: Absorber stack thicknesses comparable to the EUV wavelength necessitate rigorous modeling beyond thin-mask approximations. The convolution-based approach, validated by Born series simulations, is essential for fidelity.
  • Source polychromatism: The EUV-HL process is sensitive to illumination bandwidth. Masks are explicitly optimized for the source spectrum, mitigating chromatic aberrations that degrade high-spatial-frequency components.

Focus accuracy is incorporated during mask optimization by accounting for setup-specific tolerances, rendering the process robust to offsets up to ±100 nm.

Experimental Demonstration

Using the described framework and fabrication pipeline, the authors demonstrate EUV-HL with critical dimensions down to 40 nm on a range of non-periodic, curvilinear patterns. SEM imaging verifies pattern fidelity to target designs. Quantitative evaluation via binarized SEM contours and MSE calculations exhibits strong agreement between simulation and experiment, with MSE values ranging from 0.15 to 0.33 for 40–80 nm structures.

Contrasting prior EUV-HL achievements, which were limited to 372 nm CD on periodic elbow patterns, the methodology presented here substantially extends the resolution and design flexibility, enabling device-relevant geometries impossible with interference lithography.

Implications and Future Directions

EUV-HL uniquely complements projection EUV and EUV-IL, providing a lensless, cost-effective route to arbitrary nanopatterning at sub-50 nm resolution. This is especially pertinent for prototyping metasurfaces, photonic crystals, quantum emitter arrays, and superconducting nanowires, where arbitrary, non-periodic features are a bottleneck.

The inverse-design approach is wavelength-agnostic, establishing a foundation for adaptation to beyond-EUV (BEUV, ~3.x nm) regimes. This prospect is significant for early-stage BEUV resist screening and patterning, which current interference-based techniques cannot access.

Practical limitations involve mask scaling, resist sensitivity, and temporal coherence. The use of transmissive membrane masks does not scale directly to larger fields. Reflective mask architectures or robust membrane technologies will be necessary for wafer-scale printing. Higher-coherence sources (e.g., free-electron lasers, HHG) and improved resists (e.g., metal-oxide platforms) are expected to further enhance resolution and throughput. The integration of multi-beam mask writers would accelerate mask fabrication for larger patterns.

Conclusion

The methodology advances EUV-HL to achieve 40 nm critical dimension printing on arbitrary, non-periodic, curvilinear layouts, representing nearly an order-of-magnitude improvement over existing EUV-HL resolutions. Through rigorous modeling of physical effects, efficient inverse-design mask optimization, and direct-write mask fabrication, the work establishes EUV-HL as a flexible, accessible tool for sub-50 nm nanostructure prototyping. It offers a natural path to non-periodic BEUV nanopatterning and positions itself as a pivotal lensless complement to interference lithography for next-generation device research (2605.21430).

Paper to Video (Beta)

No one has generated a video about this paper yet.

Sign Up to Generate All Videos Subscribe on YouTube

Whiteboard

No one has generated a whiteboard explanation for this paper yet.

Sign Up to Generate

Explain it Like I'm 14

What is this paper about?

This paper shows a new way to “print” tiny shapes—tens of nanometers wide—using extreme ultraviolet (EUV) light without the giant, complicated lenses used in today’s chip factories. The team uses a computer-designed hologram mask to bend EUV light so it draws any pattern they want on a light‑sensitive material. They succeed in making features as small as 40 nanometers (about 2,000 times thinner than a human hair), which is almost 10× sharper than previous EUV holography results.

What questions were the researchers trying to answer?

  • Can we make very small, non‑repeating (non‑periodic), even curvy patterns using EUV light without expensive projection lenses?
  • Can a computer “work backward” from the pattern we want and design a hologram mask that, when lit by EUV, creates that pattern on a wafer?
  • What physics do we need to model correctly so the printed pattern matches the design, even with real‑world issues like slightly different colors of EUV light and 3D mask thickness?

How did they do it?

The big idea in simple terms

  • Imagine a stencil that doesn’t just block light but bends it in precise ways so that, a short distance away, the light draws a detailed picture. That stencil is a computer‑generated hologram (CGH) mask.
  • Instead of guessing the stencil, a computer designs it by starting from the picture you want and solving the problem backwards. This is called inverse design.

Designing the hologram mask

  • The team built a fast, physics‑based simulator that predicts how EUV light passes through and around the 3D features of the mask. In everyday terms, they treated the mask like a special “filter” that applies the same kind of blur and phase twist everywhere—a lot like using the same filter on every part of a photo. In math, this is a “shift‑invariant convolution” model.
  • They included two real‑world effects that really matter:
    • 3D mask effects (M3D): The mask isn’t paper‑thin; it has thickness compared to the EUV wavelength. That thickness changes how light bends and can’t be ignored.
    • Polychromatism: The light isn’t a single perfect “color” (wavelength); it’s a small spread around 13.5 nm. That spread can blur fine details if not designed for.
  • With these effects in the model, the computer adjusts the mask pattern step by step (like tuning knobs) until simulated prints closely match the target pattern. This “gradient‑based optimization” is the engine of the inverse design.

Making the mask

  • The mask is a thin window (an 80 nm silicon nitride membrane) coated with a 200 nm layer of HSQ, a material that both:
    • can be patterned at very high resolution using an electron beam, and
    • absorbs EUV light well enough to shape it.
  • They directly wrote the hologram pattern into the HSQ with electron‑beam lithography and used careful drying to stop tall, thin features from collapsing.

Printing with EUV light

  • They shined coherent EUV light from a synchrotron (a powerful light source) onto the mask. The diffracted light traveled a short distance in air to a wafer coated with HSQ resist (think “EUV‑sensitive film”).
  • After exposure of about 10–20 seconds, they developed the resist and checked the patterns with a scanning electron microscope (SEM).
  • They also compared the printed results with their simulations using a simple “difference score” (mean squared error) to confirm the model and the prints matched well.

What did they find, and why is it important?

  • They printed arbitrary patterns—including curvy shapes, letters, and logic‑chip‑like layouts—with critical dimensions down to 40 nm.
  • This is nearly a 10× improvement over the previous EUV holography record (372 nm), and crucially, it works for non‑repeating designs, not just periodic grids.
  • The approach needs no EUV projection lenses, which are extremely complex and expensive. That makes advanced prototyping more accessible to research labs and startups.
  • Because the design method is general, it can be adapted to even shorter wavelengths (so‑called BEUV around ~3 nm), opening a path to future ultrafine prototyping that today’s interference methods can’t do.

In short, the team combined the best of two worlds:

  • The simplicity of lensless methods (like interference lithography), and
  • The freedom to draw any pattern (like projection lithography), at resolutions useful for real device prototypes.

What could this lead to?

  • Faster, cheaper prototyping of tiny devices in photonics and quantum tech, like:
    • metasurfaces,
    • photonic crystals,
    • arrays of single‑photon emitters,
    • superconducting nanowire circuits.
  • Earlier testing of materials and designs for future lithography at even shorter wavelengths (BEUV), before big industrial tools exist.

There are still challenges—scaling up the mask area, getting even narrower light bandwidths, faster mask writing, and more sensitive resists—but none are fundamental blockers. This work shows EUV holographic lithography can already make sub‑50 nm, non‑periodic patterns with a simple, lensless setup, opening a flexible new tool for nanotechnology research and development.

View Paper Prompt  View All Prompts 

Knowledge Gaps

Knowledge gaps, limitations, and open questions

Below is a concise, actionable list of what remains missing, uncertain, or unexplored in the presented work.

  • Validity bounds of the shift‑invariant M3D convolution model: lack of rigorous benchmarking (e.g., full‑mask RCWA/FDTD on reduced fields) across pattern types, densities, and feature sizes to quantify when/where the kernel approximation breaks down.
  • Kernel representativeness and stationarity: how the single “representative” kernel handles spatially varying mask contexts (corners, isolated vs dense, curvilinear features); need for position‑dependent/adaptive kernels and criteria to select/update them.
  • Spatial partial coherence and source angular distribution: the model treats spectral incoherence, but does not explicitly quantify or validate spatial coherence/illumination NA effects (mutual intensity, angular sampling) on aerial image fidelity.
  • Sensitivity to wavefront errors and beam non‑idealities: no analysis of tolerances to phasefront aberrations, illumination inhomogeneity, or source jitter; lack of wavefront metrology to correlate with print fidelity.
  • Polychromatic optimization robustness: how mask performance degrades as bandwidth increases beyond ~4%, under bandwidth drifts, or with different source types (synchrotron vs HHG vs FEL); guidance on wavelength sampling density vs runtime.
  • Depth‑of‑focus (DOF) characterization: no systematic DOF/process window maps across diverse patterns; only a ±100 nm focus tolerance example—unclear how DOF scales with CD, pitch, or mask design choices.
  • Resolution–working‑distance–mask‑size trade‑offs: absence of a quantitative exploration/optimizer to select working distance and mask aperture for a desired CD and field‑of‑view (FOV), including practical limits (membrane size, stage clearance).
  • Sub‑feature‑size printing not yet demonstrated: the paper asserts feasibility of printing CDs below the smallest mask feature but does not show it; missing error budgets and design rules for safely operating in this regime.
  • Throughput and scalability of mask fabrication: no data on e‑beam write time, yield, proximity‑effect correction (PEC) strategy, or variability for masks larger than 2×2 mm²; unclear path to large‑area CGHs with acceptable cycle times.
  • Mechanical robustness and lifetime of transmissive membranes: unaddressed issues of membrane sag, vibration, thermal load, EUV‑induced damage/contamination, and their impact on aerial image stability over repeated use.
  • Reflective EUV holographic masks: feasibility, design methodology, and M3D modeling for multilayer reflective holograms (phase control, standing waves, flare/back‑reflection) remain open for larger FOV scaling.
  • Diffraction efficiency and background management: no measurement of hologram efficiency, zero‑order/background suppression, or stray scattering; impact on dose, image contrast, and process latitude is not quantified.
  • Resist/process metrics: absence of dose‑to‑clear curves, dose latitude, line‑edge/line‑width roughness (LER/LWR), CDU statistics, and development bias; the resist model is a simple sigmoid not calibrated to experimental kinetics.
  • Photon shot‑noise and stochastic effects: no analysis of stochastic roughness limits at the reported doses and CDs, nor their dependence on image contrast and bandwidth.
  • Overlay and multi‑layer alignment: no demonstration of alignment marks, stage control, or overlay budgets for layer‑to‑layer patterning—critical for device prototyping.
  • Field uniformity and stitching: lack of CD/uniformity maps across the 2×2 mm² membrane, assessment of image distortion, or evaluation of stitching errors for tiled exposures.
  • Sensitivity to mask fabrication deviations: no quantitative study of how CD bias, sidewall slope, roughness, or height variation perturb the aerial image; manufacturability‑aware optimization loop (including e‑beam PEC) is not integrated.
  • Absorber material/stack optimization: HSQ transmits ~13.5% at 13.5 nm; the effect of absorber transmission on image contrast is not quantified; alternatives (higher‑attenuation, phase‑shifting, or hybrid amplitude‑phase masks) are unexplored.
  • Phase control on mask: only binary‑thickness (amplitude‑dominant) masks are reported; potential gains from engineered phase (e.g., membrane thickness modulation, multi‑level structures) are unaddressed.
  • Realistic wafer stacks: imaging over industrial EUV stacks (underlayers, reflectivity, standing waves) and topography is not evaluated; current results are limited to HSQ on flat substrates.
  • Mechanical/thermal stability during 10–20 s exposures: no quantified vibration/drift tolerances or environmental sensitivity; implications for longer exposures or less stable sources remain unknown.
  • Algorithm scaling and convergence: no absolute runtimes/memory footprints, convergence statistics, or failure modes; scalability to cm‑scale masks and high‑density layouts remains unproven.
  • Loss‑function design and trade‑offs: limited detail on weighting between resolution, DOF, background suppression, and robustness; no sensitivity analysis on hyperparameters or multi‑objective optimization strategies.
  • Material optical constants and uncertainties: impact of n,k uncertainties (HSQ, Si₃N₄) across the source bandwidth on the M3D kernel and aerial image is not quantified; need for in‑situ optical metrology.
  • Path to BEUV (~3.x nm): feasibility lacks quantitative analysis of material absorption, membrane viability, reflective implementation, and source coherence; no experimental validation or end‑to‑end design study.
  • Illumination conditioning requirements: unspecified needs for bandwidth filtering, spatial filtering, or coherence control for HHG/FEL sources to achieve comparable fidelity; practical instrument designs are absent.
  • Metrology methodology: reliance on binarized SEM contours and MSE introduces threshold bias; lacks edge placement error (EPE), contour fidelity metrics, and correlation to simulated latent images.
  • Dose–resolution trade‑offs: no systematic mapping of CD vs dose vs bandwidth vs working distance to guide design under practical dose constraints.
  • Pattern‑class performance: limited to a few glyphs and elbows; no broad evaluation across isolated, dense, and curvilinear patterns with varying pitches and CDs to define design rules.
  • Working‑distance tolerances: sensitivity of image fidelity to deviations from the 200 µm working distance (e.g., wafer bow, membrane–wafer spacing errors) is not quantified.
  • Stray light from supports and tooling: potential background from the Si frame, membrane edges, or holder reflections not measured or mitigated.
  • Robustness of inverse design: lack of reporting on cases where optimization stagnates or fails, dependence on initialization/noise seeding, and strategies to avoid local minima.
View Paper Prompt  View All Prompts 

Practical Applications

Immediate Applications

The following applications can be deployed now using the demonstrated 40 nm EUV holographic lithography (EUV-HL), the inverse-design CGH framework, and the HSQ-based mask/resist workflow.

  • Arbitrary-pattern EUV nanostructure prototyping at synchrotron beamlines [Sectors: semiconductors, photonics, quantum, materials]
    • Use cases: rapid iteration of non-periodic, curvilinear features for metasurfaces, photonic crystals, superconducting nanowire meanders (SNSPDs), single-photon emitter arrays, IC logic-layer-like test patterns.
    • Tools/products/workflows: inverse-design CGH software (with M3D convolution and polychromatic optimization), direct-write HSQ transmissive holographic masks, adapted EUV-IL end stations for HL exposures, SEM-based verification with simulation overlay.
    • Assumptions/dependencies: access to a coherent EUV source (e.g., synchrotron); e-beam capability to pattern HSQ at 40 nm; small field of view and 200 µm working distance; CPD for mask development; exposure stability and focus uncertainty accounted for by focus-tolerant optimization.
  • Arbitrary-pattern EUV resist screening and process-window exploration [Sectors: semiconductors, tool vendors]
    • Use cases: evaluate resist sensitivity, LER/LWR, CD control on non-periodic and curvilinear geometries not addressable by EUV interference lithography (EUV-IL). Enables DTCO studies with realistic shapes.
    • Tools/products/workflows: curated HL test-pattern libraries; automated MSE-based image-to-design comparison; dose–focus matrix using focus-tolerant masks.
    • Assumptions/dependencies: current demonstration uses HSQ (longer doses vs HVM requirements); bandwidth (~4% FWHM) and partial coherence must be included in design; beamtime availability.
  • Rapid feasibility checks for curvilinear DTCO and inverse lithography concepts [Sectors: EDA/software, semiconductors]
    • Use cases: evaluate manufacturability of curvilinear layouts (e.g., elbows, glyphs) with a lensless route; inform OPC/ILT strategies with empirical aerial-image and resist outcomes.
    • Tools/products/workflows: CGH optimizer integrated with layout tools; design–simulation–exposure–SEM loop; M3D vs thin-mask modeling comparisons for education/validation.
    • Assumptions/dependencies: physics differences vs projection EUV limit direct transferability; requires synchronization of layout grid and mask pixelation (40 nm used here).
  • Beamline-accessible “low-cost EUV” path for small labs and startups [Sectors: academia, startups/SMEs]
    • Use cases: device prototyping without access to projection EUV scanners; shared-facility model at national labs for advanced nanofabrication.
    • Tools/products/workflows: facility-provided mask-writing service (HSQ-on-Si3N4), HL exposure bookings, standard operating procedures for dose calibration and focus bracketing.
    • Assumptions/dependencies: scheduling on shared EUV beamlines; mechanical robustness of 2 × 2 mm² membranes; limited wafer-scale throughput.
  • Aerial-image engineering and diagnostics [Sectors: EUV/X-ray optics, metrology]
    • Use cases: prescribe and generate complex intensity distributions for optical diagnostics, alignment marks, or instrument calibration; fabricate challenging curvilinear metrology targets.
    • Tools/products/workflows: CGH patterns for beam-shaping tests; SEM and aerial-image contour overlays; structured targets for SEM/LER benchmarking.
    • Assumptions/dependencies: indirect aerial-image assessment through resist patterns; sensitivity to source bandwidth and stability.
  • Education and training in EUV imaging physics [Sectors: academia, workforce development]
    • Use cases: hands-on curricula contrasting thin-mask vs M3D physics, polychromatic effects, angular-spectrum propagation, and focus-tolerant design.
    • Tools/products/workflows: open or shared simulation scripts; teaching masks with known ground-truth aerial images and SEM outcomes.
    • Assumptions/dependencies: compute resources for optimization; access to beamline time for lab courses or demonstrations.
  • Early validation of non-periodic test vehicles for materials and devices [Sectors: materials science, sensors]
    • Use cases: templating arrays (e.g., catalyst, quantum-dot nucleation sites), nanoscale gaps and bends for plasmonics or nanomagnetics, non-periodic nanofluidic features.
    • Tools/products/workflows: HL-generated masks tailored to materials experiments; short exposure times (10–20 s reported) allow quick turnaround for small fields.
    • Assumptions/dependencies: materials compatibility with HSQ/e-beam process; sample handling under cleanroom conditions; small-area limits.

Long-Term Applications

These opportunities require further research, scaling, new materials/sources, or engineering development before practical deployment.

  • BEUV (≈3.x nm) non-periodic pattern prototyping and resist/process development [Sectors: semiconductors]
    • Use cases: early-stage BEUV resist screening and layout exploration where EUV-IL is inapplicable; pathfinding for BEUV DTCO.
    • Tools/products/workflows: CGH framework extended to BEUV; adapted materials stacks; exposure on FEL/HHG or advanced BEUV sources.
    • Assumptions/dependencies: availability of BEUV sources and optics; absorber/membrane materials with suitable BEUV properties; increased sensitivity to polychromatism.
  • Scaling EUV-HL to larger fields of view and higher throughput [Sectors: semiconductors, equipment]
    • Use cases: pre-production patterns and larger prototypes beyond mm² areas; stitched exposures for die-scale layouts.
    • Tools/products/workflows: reflective HL masks or reinforced large-area membranes; precision scanning stages and stitching algorithms; multi-exposure tiling; in-situ focus control.
    • Assumptions/dependencies: mechanical robustness, alignment accuracy, vibration control; multi-beam mask writers to reduce mask write time.
  • Tabletop EUV-HL instruments [Sectors: instrumentation, SMEs, R&D services]
    • Use cases: in-house lensless EUV patterning for labs without synchrotron access.
    • Tools/products/workflows: HHG-based EUV sources; compact exposure stages with environmental isolation; integrated CGH design software.
    • Assumptions/dependencies: sufficient source brightness/coherence; improved resists to keep doses practical; safety and contamination control.
  • Commercialization of the CGH inverse-design stack [Sectors: EDA/software, computational lithography]
    • Use cases: “Holographic OPC” for HL; cloud/HPC services for mask synthesis; plug-ins for standard CAD flows with curvilinear support.
    • Tools/products/workflows: production-grade M3D convolution engines, Born-series-based kernels, polychromatic/focus-tolerant loss functions, verification suites (MSE, contour overlays).
    • Assumptions/dependencies: model calibration across processes and sources; IP/licensing; integration with layout standards and PDKs.
  • Sub-20 nm curvilinear patterning by phase–amplitude holograms and advanced resists [Sectors: semiconductors, photonics]
    • Use cases: tighter CDs and smoother contours for high-performance photonics and quantum devices.
    • Tools/products/workflows: multi-level or phase-shifting holographic masks; metal-oxide/inorganic resists with higher EUV sensitivity and lower LER; higher temporal coherence sources (FEL/HHG).
    • Assumptions/dependencies: fabrication of multi-level EUV masks; source coherence improvements; dose budgets and process control.
  • Depth-of-focus engineering and multi-plane holography [Sectors: photonics, materials]
    • Use cases: robust printing across focus errors; controlled multi-depth exposures for quasi-3D patterning via layerwise or multi-pass strategies.
    • Tools/products/workflows: multi-plane loss functions; adaptive focusing and stage control; resist/process stacks tuned for sequential exposures.
    • Assumptions/dependencies: precise stage metrology; process stability across multiple doses; resist chemistry enabling multi-exposure latitude.
  • Targeted device pipelines in quantum and photonics [Sectors: quantum sensing/communications, integrated photonics]
    • Use cases: faster prototyping-to-product cycles for SNSPDs, waveguide couplers, nanophotonic cavities, emitter arrays with arbitrary placements.
    • Tools/products/workflows: design kits linking optical/electrical specs to HL-ready patterns; validated fabrication recipes; reliability and packaging flows.
    • Assumptions/dependencies: transferability from small-field HL prototypes to scalable manufacturing (projection EUV or alternative techniques), or maturation of HL-scale-up approaches.
  • Metrology and standards for EUV/BEUV patterning [Sectors: standards bodies, metrology vendors]
    • Use cases: canonical non-periodic test artifacts for SEM/LER calibration, proximity-effect studies, aerial-image validation at EUV/BEUV.
    • Tools/products/workflows: certified HL masks and wafers; traceable measurement protocols; repositories of target/aerial/SEM datasets.
    • Assumptions/dependencies: cross-tool repeatability; stability of membrane-based artifacts; community adoption.
  • Policy and infrastructure enabling broader access [Sectors: policy, funding agencies]
    • Use cases: democratize EUV/BEUV R&D through shared HL capability and mask-writing services; support regional hubs.
    • Tools/products/workflows: facility programs (beamtime, mask fabrication queues), open pattern libraries, training grants.
    • Assumptions/dependencies: sustained funding; coordination among national labs and universities; export-control compliance.
  • Indirect consumer impacts via accelerated device innovation [Sectors: daily life, broader economy]
    • Use cases: faster iteration cycles for advanced ICs, sensors, and photonic components benefiting computing, communications, and imaging.
    • Tools/products/workflows: tighter feedback between design and lab-scale validation; earlier risk retirement for novel layouts.
    • Assumptions/dependencies: effective tech transfer from HL prototypes to scalable manufacturing; ecosystem adoption of HL-informed DTCO.
View Paper Prompt  View All Prompts 

Glossary

  • aberration: Optical imperfections causing image blur or distortion. "flare, aberration, and wavefront distortion"
  • absorber: The mask material or layer that attenuates EUV light to shape the diffracted field. "patterned by electron-beam lithography to form the absorber."
  • aerial image: The light intensity distribution formed at the wafer plane before resist processing. "where its intensity distribution reproduces the target aerial image."
  • angular-spectrum method: A Fourier-domain wave-propagation technique used to compute fields from mask to wafer. "computed using the angular-spectrum method."
  • aspect ratio: The height-to-width ratio of a fabricated feature. "The pixel size of 40 nm creates features with aspect ratio of 5:1"
  • BEUV (beyond-EUV): A shorter-wavelength regime (~3.x nm) targeted for future lithography. "beyond-EUV (BEUV) wavelengths"
  • binarization: Converting a grayscale latent image into a binary resist pattern via thresholding. "represents the binarization of the latent image"
  • Born series: An iterative approximation for solving wave equations in scattering problems. "simulated based on Born series, a fast numerical method for solving Helmholtz equation"
  • capillary forces: Surface-tension forces in liquids that can collapse high-aspect-ratio structures during drying. "increasing the risk of pattern collapse due to aqueous capillary forces"
  • chromatic aberration: Image degradation arising from wavelength-dependent focusing/phase effects. "because chromatic aberration predominantly degrades the high-spatial-frequency components of the aerial image"
  • computer-generated hologram (CGH): A numerically designed hologram pattern that produces a target image upon illumination. "a computer-generated hologram (CGH) mask is illuminated with coherent EUV light"
  • convolution (shift-invariant): Modeling the mask’s local 3D scattering by convolving a position-independent kernel with the mask layout. "a shift-invariant convolution model"
  • critical dimension (CD): The key feature width in a pattern, used as a resolution metric. "with a critical dimension (CD) of 70 nm."
  • critical point drying (CPD): A drying process that avoids capillary collapse by passing through the critical point of a fluid. "necessitating the use of critical point drying (CPD) after resist development."
  • curvilinear: Patterns with smoothly varying, non-Manhattan geometries. "arbitrary, non-periodic, curvilinear patterning"
  • defocus: Displacement from the optimal focal plane that degrades image fidelity. "MSE as a function of defocus"
  • depth of focus: The range of focus positions over which image quality remains acceptable. "to enlarge the depth of focus"
  • diffraction: The bending and interference of waves due to structures, used here for lensless imaging. "By utilizing diffraction-based imaging"
  • electron-beam lithography: Direct-writing technique using a focused electron beam to pattern resist. "direct-write electron-beam lithography in hydrogen silsesquioxane"
  • extreme ultraviolet (EUV): Short-wavelength radiation around 13.5 nm used for advanced lithography. "Extreme ultraviolet (EUV) lithography is the cornerstone of the fabrication of advanced integrated circuits"
  • EUV holographic lithography (EUV-HL): Lensless lithography using CGH masks to create arbitrary patterns at EUV wavelengths. "EUV holographic lithography (EUV-HL) as a lensless route to arbi- trary, non-periodic, curvilinear patterning"
  • EUV interference lithography (EUV-IL): Lensless lithography that forms periodic patterns by interfering coherent EUV beams. "EUV interference lithography (EUV-IL) provides a lensless alternative"
  • FDTD (finite-difference time-domain): A numerical method for solving Maxwell’s equations in time domain. "A rigorous NF calculation, which would require a full 3D electromag- netic simulation (FDTD or RCWA)"
  • field of view (FOV): The lateral area on the wafer intended to be patterned in one exposure. "To print a pattern over a given field of view FOV"
  • flare: Stray light in optical systems that reduces image contrast. "operate with minimal flare, aberration, and wavefront distortion"
  • Fresnel zone plate: A diffractive lens with concentric zones that focuses light by diffraction. "the holographic mask that generates it is a Fresnel zone plate"
  • FWHM (full width at half maximum): A measure of spectral or spatial width at half the peak value. "spectral bandwidth of approximately 4% (FWHM)"
  • Helmholtz equation: A fundamental wave equation used to model steady-state fields. "a fast numerical method for solving Helmholtz equation"
  • high-order harmonic generation (HHG): A coherent EUV/X-ray source produced by nonlinear upconversion of laser light. "high-order harmonic generation source"
  • hydrogen silsesquioxane (HSQ): A negative-tone inorganic resist also used here as an EUV absorber. "hydrogen silsesquioxane (HSQ) that was patterned by electron-beam lithography"
  • incoherent summation: Intensity addition of independently phased (mutually incoherent) field components. "The total intensity distribution at the wafer is obtained as the incoherent summation of all monochromatic aerial-image components"
  • inverse-design: Optimization-driven design process that computes mask patterns from target images via a forward model. "We introduce an inverse-design framework"
  • line-edge roughness: Random deviations along a feature edge that degrade pattern fidelity. "like line-edge roughness, CD bias, or sidewall slope"
  • Manhattan-geometry: Orthogonal (axis-aligned) layout style typical of IC designs. "a random Manhattan-geometry layout"
  • mask three-dimensional (M3D) effects: 3D electromagnetic interactions in thick masks that modify transmitted fields beyond thin-mask assumptions. "we will quantify mask three-dimensional (M3D) effects"
  • mean square error (MSE): A quantitative metric of image or pattern deviation used as a loss function. "MSE as a function of defocus"
  • metasurfaces: Planar arrays of subwavelength structures that control wavefronts. "metasurfaces, photonic crystals, quantum emitter arrays, and superconducting nanowire structures"
  • near field (NF): The complex optical field immediately after the mask before free-space propagation. "a corresponding near field (NF) is generated at the object plane"
  • numerical aperture: A dimensionless parameter setting resolution by limiting spatial frequencies. "analogous to how the numerical aperture of a projection system determines the resolution of the aerial image."
  • paraxial: Approximation assuming small angles relative to the optical axis. "In the idealized monochromatic paraxial case"
  • photon sieve: A diffractive focusing element using a distribution of pinholes to replace continuous Fresnel zones. "A photon sieve, in which the continuous Fresnel zones are replaced by discrete pinholes"
  • polychromatism: Presence of a finite spectral bandwidth rather than a single wavelength. "M3D scattering and source polychromatism"
  • projection optics: The imaging lens system in scanners that projects the mask pattern onto the wafer. "the projection optics: a multi-element reflective lens assembly"
  • RCWA (rigorous coupled-wave analysis): A frequency-domain solver for periodic structures in electromagnetics. "a full 3D electromag- netic simulation (FDTD or RCWA)"
  • reflective projection optics: Mirror-based imaging optics required at EUV wavelengths. "multi-element reflective projection optics"
  • resist: A photosensitive material whose solubility changes upon exposure and development. "we used HSQ as the imaging resist on the wafer"
  • shift-invariant: System property where the response is the same regardless of lateral position. "the mask three-dimensional (M3D) response as a shift-invariant convolution"
  • sigmoid function: A smooth S-shaped function used here to model the resist thresholding response. "the resist response is modeled by a sigmoid function"
  • synchrotron: A high-brightness radiation source producing coherent EUV used for exposures. "expose them with synchrotron-generated EUV radiation"
  • thin-mask approximation: Simplified model treating mask regions as purely transparent or opaque without 3D scattering. "The thin-mask approximation, which assigns NF amplitudes of 1 and 0"
  • wafer: The substrate onto which patterns are transferred. "the wafer plane"
  • wavefront distortion: Deviation of the optical wavefront from ideal shape, degrading imaging. "flare, aberration, and wavefront distortion"
  • working distance: Separation between the mask and wafer during exposure. "the working distance is 200 um"
View Paper Prompt  View All Prompts 

Open Problems

We haven't generated a list of open problems mentioned in this paper yet.

Generate Now

Collections

Sign up for free to add this paper to one or more collections.

Sign Up

Tweets

Sign up for free to view the 1 tweet with 117 likes about this paper.

Sign Up for Free