Neural field enhanced phase retrieval of atomic-scale structural dynamics in radiation sensitive materials

Area of Science:

• Materials Science
• Life Science
• Microscopy

Background:

• Atomic-scale imaging of radiation-sensitive materials is challenging due to radiation damage.
• Low-dose transmission electron microscopy (TEM) minimizes damage but yields noisy, low-resolution images.
• Fine structure analysis of sensitive materials requires improved imaging techniques.

Purpose of the Study:

• To present a novel phase retrieval method for high-quality atomic-scale imaging of radiation-sensitive materials.
• To overcome the limitations of low-dose TEM in analyzing delicate structures.
• To enable detailed analysis of materials like halide perovskites under electron irradiation.

Main Methods:

• Integration of neural fields (NF) with traditional exit wave reconstruction (EWR).
• Application of the EWR-NF method to low-dose TEM imaging.
• Reconstruction of atomic structures from limited low-dose experimental data.

Main Results:

• The EWR-NF method successfully revealed atomic details from minimal low-dose TEM data.
• Superior performance in reconstructing the pristine atomic structure of methylammonium lead iodide (MAPbI3).
• Enabled higher temporal resolution imaging of irradiation-induced decomposition intermediates.

Conclusions:

• EWR-NF offers a powerful tool for atomic-level structure analysis of sensitive halide perovskites.
• The method facilitates understanding of irradiation-induced structure changes in materials.
• Broad implications for materials science and related fields requiring sensitive sample analysis.