EUV_FPM

Enhanced EUV Mask Imaging Using Fourier Ptychographic Microscopy

Presented at SPIE Optical and EUV Nanolithography XXXVIII, February 26, 2025.

Preprint manuscript and example datasets are available upon reasonable request. Contact: [email protected]

Basic Usage

Single-Patch FP Reconstruction

Run recon.py with a selected configuration file from configs/recon/:

1. Choose a configuration file in configs/recon/.
2. Set the config_name parameter in recon.py to the target configuration file name.

Full-FOV FP Reconstruction (with Automatic Patch Division)

Run recon_full.py for a full-FOV reconstruction using automatically divided patches:

1. Choose a configuration file in configs/recon_full/.
2. Set the config_name parameter in recon_full.py to the target configuration file name.

Reconstruction visualization and refinement

Check the plot_exp.ipynb and torch_fp_recon.ipynb and choose the configuration file accordingly.

Configuration File Key Fields

Below are a few key parameters in the configuration files:

• elliptical_pupil:

  • Set to True for anamorphic 0.55/4× NA with 20% central obscuration and 1250× magnification.
  • Set to False for isomorphic 0.33/4× NA with 900× magnification.

• ROI_length:

  • Defines the reconstruction patch size.

• algorithm:

  • Specifies the reconstruction algorithm. Choose from:
    • 'GS' – Gerchberg-Saxton
    • 'GN' – Gauss-Newton
    • 'EPFR' – Embedded Pupil Function Recovery

Full-FOV Reconstruction Demo

Data collected at 13.5 nm wavelength EUV microscope.

(a) Details of three sample regions in the full FOV, showing one raw captured image and the corresponding reconstruction for each. The gray box in the bottom left corner marks the patch used for field-dependent aberration analysis in panel (b). The gray dots indicate the centers of the segmented regions, with the full FOV divided into a 6×6 grid with 20% overlap.
(b) Full-FOV aberration visualization, where the 6×6 grid layout presents the pupil wavefront reconstruction for each patch.