Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Tabletop Lensless Imaging in the Extreme Ultraviolet with Reduced Radiation Dose

Sukyoon Oh, Monalisa Mallick, Thomas Siefke, Christian Spielmann

TL;DR

The paper demonstrates a lensless, tabletop XUV imaging platform that uses high-harmonic generation and correlation-based ghost imaging to achieve micrometer-scale imaging with dramatically reduced radiation dose. By systematically evaluating illumination patterns (Hadamard, differential Hadamard, Fourier, and random) and reconstruction methods (traditional GI and compressive sensing), the authors show Hadamard-based strategies offer superior reconstruction quality, especially when combined with CS. A data-processing pipeline—encompassing pattern calibration, real-pattern usage, histogram equalization, and noise-cropped bucket data—yields up to a ~400% improvement in image similarity metrics over baseline conditions, validating a robust, low-dose approach for non-destructive inspection of delicate materials. The work broadens access to XUV microscopy, highlighting practical pathways toward compact, optics-free, dose-efficient imaging suitable for materials science and biology.

Abstract

High-resolution extreme ultraviolet (XUV) imaging remains limited by conventional approaches that require complex optics such as multilayer mirrors and zone plates. These methods are expensive, suffer from chromatic aberrations and narrow fields of view, and demand highly stable, coherent beam sources typically found only at large-scale facilities. Critically, the high photon flux they require often damages sensitive biological and soft-matter samples. We present a new solution: a lensless XUV microscopy platform combining a compact tabletop high-harmonic generation source with correlation-based ghost imaging. Our approach eliminates the need for complex optics, lowering system cost and dramatically improving resilience against lab-scale instabilities. Leveraging Hadamard patterns and compressive sensing algorithms, we achieve high-fidelity imaging even in low-photon environments, with a 400\% improvement in structural similarity index compared to baseline methods. This confirms the feasibility of broadband, low-dose XUV imaging, enabling damage-minimized, non-destructive inspection for advanced materials and biological specimens, and establishes a new paradigm for accessible XUV microscopy.

Tabletop Lensless Imaging in the Extreme Ultraviolet with Reduced Radiation Dose

TL;DR

The paper demonstrates a lensless, tabletop XUV imaging platform that uses high-harmonic generation and correlation-based ghost imaging to achieve micrometer-scale imaging with dramatically reduced radiation dose. By systematically evaluating illumination patterns (Hadamard, differential Hadamard, Fourier, and random) and reconstruction methods (traditional GI and compressive sensing), the authors show Hadamard-based strategies offer superior reconstruction quality, especially when combined with CS. A data-processing pipeline—encompassing pattern calibration, real-pattern usage, histogram equalization, and noise-cropped bucket data—yields up to a ~400% improvement in image similarity metrics over baseline conditions, validating a robust, low-dose approach for non-destructive inspection of delicate materials. The work broadens access to XUV microscopy, highlighting practical pathways toward compact, optics-free, dose-efficient imaging suitable for materials science and biology.

Abstract

High-resolution extreme ultraviolet (XUV) imaging remains limited by conventional approaches that require complex optics such as multilayer mirrors and zone plates. These methods are expensive, suffer from chromatic aberrations and narrow fields of view, and demand highly stable, coherent beam sources typically found only at large-scale facilities. Critically, the high photon flux they require often damages sensitive biological and soft-matter samples. We present a new solution: a lensless XUV microscopy platform combining a compact tabletop high-harmonic generation source with correlation-based ghost imaging. Our approach eliminates the need for complex optics, lowering system cost and dramatically improving resilience against lab-scale instabilities. Leveraging Hadamard patterns and compressive sensing algorithms, we achieve high-fidelity imaging even in low-photon environments, with a 400\% improvement in structural similarity index compared to baseline methods. This confirms the feasibility of broadband, low-dose XUV imaging, enabling damage-minimized, non-destructive inspection for advanced materials and biological specimens, and establishes a new paradigm for accessible XUV microscopy.
Paper Structure (15 sections, 5 equations, 6 figures, 1 table)

This paper contains 15 sections, 5 equations, 6 figures, 1 table.

Table of Contents

  1. Introduction
  2. Materials and Methods
  3. Experimental Setup
  4. XUV Mask Design and Fabrication
  5. Ghost Imaging Reconstruction Algorithm and Patterns
  6. Fundamentals of Traditional Ghost Imaging Algorithm
  7. Compressive Sensing: Beyond Traditional Methods
  8. Reconstruction via Differential Hadamard Patterns
  9. Fourier Patterns: Reconstruction Algorithm and Advantages
  10. Performance of Various Patterns in Ghost Imaging Simulation
  11. Results
  12. XUV Ghost Imaging: Experimental Results
  13. Improving Image Quality through a Data Processing Pipeline
  14. Image Quality Optimization
  15. Discussion and Future Research

Figures (6)

  • Figure 1: XUV Ghost Imaging Experimental Setup. (a) A schematic of the experimental setup, from high harmonic generation to image reconstruction. (b) The measured XUV beam profile. (c) The XUV spectrum, illustrating the effect of the XUV mask on photon counts.
  • Figure 2: The Manufactured XUV Mask. This figure presents the design and characteristics of the manufactured XUV mask. (a) A photograph of the mask, showing four sections with distinct positive Hadamard, negative Hadamard, 4-step Fourier, and random patterns. (b) A cross-sectional view illustrating the material layers. (c) A visualization of the patterns as illuminated by the XUV beam, highlighting manufacturing imperfections.
  • Figure 3: Ghost Imaging Simulation with Various Patterns. (a) Original object. (b)-(c) Reconstructions using a random pattern with (b) TGI method and (c) CS. (d) Reconstruction using a binary 4-step Fourier pattern. (e)-(g) Reconstructions using Hadamard patterns with (e) inverse transform and (f)-(g) CS.
  • Figure 4: Ghost Imaging Reconstructions of a Cross-Shaped Target. This figure compares ghost imaging reconstructions from a single experiment, showing that structured patterns successfully reconstructed the target where random patterns failed. (a) Original target. (b)-(c) Reconstructions with random patterns using TGI and CS. (d) A successful reconstruction with a differential Hadamard pattern. (e)-(g) Reconstructions using Hadamard patterns with (e) a negative and (f) a positive Hadamard Pattern, and (g) from Wafer 1.
  • Figure 5: Data Processing Pipeline for Ghost Image Reconstruction. The figure illustrates the key steps in processing patterns and bucket data to enhance ghost imaging quality. (a) Reference target. Panels (b--e) show the evolution from ideal to a processed real pattern. Panels (f--g) show the full and cropped bucket data, demonstrating noise reduction through signal isolation.
  • ...and 1 more figures