arXiv subjects

Benjamin Colmey

Publications and source records attributed to Benjamin Colmey.

4 recordsLinked to original sources

arXiv2026

The role of absorption in three-dimensional electron diffraction dynamical structure refinement

The role of absorption in 3D electron diffraction is established through analytical theory, simulation, and dynamical refinement. A two-beam expression for the absorbed integrated intensity in centrosymmetric crystals is derived, showing that for $t/\xi_g \ll 1$ reflections follow a uniform exponential decay set by the mean absorptive potential $U_0'$. Many-beam simulations of both centrosymmetric and non-centrosymmetric crystals reveal additional reflection-specific anomalous absorption beyond the uniform attenuation set by $U_0'$. Neglecting these effects in dynamical refinement of integrated intensities incurs an error that increases approximately linearly with thickness, with this error becoming more severe near zone axes. Dynamical refinements were performed on CsPbBr$_3$, quartz, and borane, with the inclusion of absorption yielding an improvement in $R_{\mathrm{obs}}$ from $6.4$ to $5.3$ \% for CsPbBr$_3$ and negligible improvements for quartz and borane. Anomalous absorption may therefore be ignored for routine refinement of integrated intensities except in high-$Z$ materials at thicknesses approaching $\xi_g $.

cond-mat.mtrl-sci↗

arXiv2025

Hybrid Physics-Machine Learning Models for Quantitative Electron Diffraction Refinements

High-fidelity electron microscopy simulations required for quantitative crystal structure refinements face a fundamental challenge: while physical interactions are well-described theoretically, real-world experimental effects are challenging to model analytically. To address this gap, we present a novel hybrid physics-machine learning framework that integrates differentiable physical simulations with neural networks. By leveraging automatic differentiation throughout the simulation pipeline, our method enables gradient-based joint optimization of physical parameters and neural network components representing experimental variables, offering superior scalability compared to traditional second-order methods. We demonstrate this framework through application to three-dimensional electron diffraction (3D-ED) structure refinement, where our approach learns complex thickness distributions directly from diffraction data rather than relying on simplified geometric models. This method achieves state-of-the-art refinement performance across synthetic and experimental datasets, recovering atomic positions, thermal displacements, and thickness profiles with high fidelity. The modular architecture proposed can naturally be extended to accommodate additional physical phenomena and extended to other electron microscopy techniques. This establishes differentiable hybrid modeling as a powerful new paradigm for quantitative electron microscopy, where experimental complexities have historically limited analysis.

physics.comp-ph↗

arXiv2025

Partial Differential Operators and Fourier Integral Operators in the Gevrey setting and applications

We state and prove here semiclassical results about the construction of asymptotic solutions by the WKB method for pseudo-differential equations of real principal type. It is a Gevrey version; the smooth $C^\infty$ and the analytic ones may be found in H\"ormander (Hormander1985) and Sj\"ostrand (Sj\"ostrand1982). We present here three versions depending on the degree of entrance in the complex domain.

math.AP↗

arXiv2024

Sub-cycle Nanotip Field Emission of Electrons Driven by Air Plasma Generated THz Pulses

Terahertz pulses generated by two-color laser plasmas have reported peak field strengths exceeding MV/cm, and when illuminating metal nanotips the near-field enhancement at the tip apex should result in extremely high bunch charges and electron energies via sub-cycle cold field emission. Here, electron emission from tungsten nanotips driven by THz pulses generated by a long filament air-plasma are reported. Electron energies up to 1.1 keV and bunch charges up to 2x$10^5$ electrons per pulse were detected, well below values expected for peak field calculated via the time averaged Poynting vector. Investigations revealed a failure in the use of the time-averaged Poynting vector when applied to long filament THz pulses, due to spatio-temporal restructuring of the THz pulse in the focus. Accounting for this restructuring significantly reduces the field strength to approximately 160 ~kV/cm, consistent with the observed electron bunch charges, peak energies and their dependence on the tip position in the THz focus. Despite these findings, our results surpass previous THz plasma-driven electron generation by an order of magnitude in both electron energy and bunch charge and a path to increasing these by an additional order of magnitude by modification of the THz optics is proposed.

cond-mat.mtrl-sci↗