Understanding phonon selection and interference in momentum-resolved electron energy loss spectroscopy

Thomas W. Pfeifer, Harrison A. Walker, Henry Aller, Samuel Graham, Sokrates Pantelides, Jordan A. Hachtel, Patrick E. Hopkins, Eric R. Hoglund · 2025

Read Understanding phonon selection and interference in momentum-resolved electron energy loss spectroscopy by Thomas W. Pfeifer, Harrison A. Walker, Henry Aller, Samuel Graham, Sokrates Pantelides, Jordan A. Hachtel, Patrick E. Hopkins, Eric R. Hoglund (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-09792.txt.

Introduction

Thomas Pfeifer, Sokrates Pantelides, Jordan Hachtel, Patrick Hopkins, Eric Hoglund. Co-workers clarified how phonons are selected in a microscope technique in a 2025 paper. Momentum-resolved tiny charged bit energy loss spectroscopy gauges lattice vibrations, phonons, with fine spatial detail. The signal mixes selection rules and interference in ways users can misread. Some phonon branches are forbidden by bouncing rules in certain geometries, while simultaneous sampling of all basis atoms forces a coherent treatment. Producing an effective zone larger than the standard Brillouin zone, which the authors name the interferometric Brillouin zone.

Their new formalism for waves and vibrations grounds the picture mathematically. Math work from molecular motion and lattice motion are checked against set up mock run methods on well-studied stuff. Including the use of selection rules to extract a polarization-selective vibrational spectrum.

Catalog record

CREATORThomas W. Pfeifer, Harrison A. Walker, Henry Aller, Samuel Graham, Sokrates Pantelides, Jordan A. Hachtel, Patrick E. Hopkins, Eric R. Hoglund
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE47 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2503-09792

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Understanding phonon selection and interference in momentum-resolved electron energy loss spectroscopy Preprint: JRN/123-ABC Thomas W. Pfeifer, Harrison A. Walker, Henry Aller, Samuel Graham, Sokrates Pantelides, Jordan A. Hachtel, Patrick E. Hopkins, Eric R. Hoglund Abstract As momentum-resolved Electron Energy Loss Spectroscopy (q-EELS) becomes more widely used for phonon measurements, better understanding of the intricacies of the acquired signal becomes necessary. Selection rules limit the allowed scattering directions, which may prohibit the appearance of specific phonon branches, and features may appear with no clear explanation. To explore both, we perform frequency-resolved q-EELS simulations via the multislice method and compare results to Spectral Energy Density (SED) and Lattice Dynamics (LD) calculations. We discuss common misconceptions of phonon behavior and characterization generally and in the context of measured or simulated q-EELS signals. We highlight the presence of an “interferometric Brillouin zone” (closely related to the Dynamic Structure Factor), resulting from the simultaneous sampling of the lattice across each basis index. We also demonstrate the scattering selection rules on well-studied systems and explore the acquisition of a polarization-selective signal based on this phenomena. Finally, we identify thickness-dependent sensitivity effects, where the upper surface of a film is sampled at a disproportionate level. We show that this can yield sensitivity to through-plane vibrational modes, which may be measured experimentally assuming sufficient sensitivity can be achieved. Our work should serve as a road map to guide momentum-resolved vibrational EELS measurements in the future, particularly as ever-increasing energy and momentum resolution becomes available. I Introduction The advent of high energy resolution in Electron Energy Loss Spectroscopy (EELS) has enabled measurements of phonon spectra in the electron microscope. This relatively recent development allows for EELS measurements at high spatial and/or momentum resolution, making the technique promising for a range of applications, including topologically isolated phonons, chiral phonons, as well as measurements in low-symmetry directions, with applications in phonon engineering. Recent works have developed the computational tools for simulating these experiments Zeiger2021 ; CR2025 . Several aspects of the measured or simulated signals remain to be addressed however. For example, selection rules Nicholls2019 may prohibit the appearance of certain phonon branches. These phenomena exist separately from issues of instrument resolution or sensitivity. Similarly, the ability to measure a localized polarization-dependent vibrational density of states (v-DOS) can lend insights into material and vibrational properties Hoglund2024 ; Yang2024 ; Li2024 , but a thorough understanding of the vibrational momentum-resolved EELS (q-EELS) signal is required. In this work, we seek a thorough explanation of all features in q-EELS signals by performing simulations in a manner similar to Zeiger Zeiger2021 and Castellanos-Reyes CR2025 , and compare results to Spectral Energy Density (SED) Thomas2010 and Lattice Dynamics (LD) calculations. These simulations and comparisons allow differentiation between effects due to the fundamental behavior of phonons versus those related to the fast-electron phonon scattering interactions. We explore two popular and commercially-relevant material systems: silicon, and aluminum nitride (AlN). Silicon is well-studied and its face-center-cubic crystal lattice is highly symmetric, reducing the number of reciprocal directions required for analysis and simplifying the simulations. Conversely, AlN has the highly anisotropic wurtzite crystal structure, enabling exploration of some of the direction-selectivity behavior of q-EELS. We believe this work will serve as a roadmap to guide future momentum-resolved vibrational EELS measurements, particularly as ever increasing energy and momentum resolution becomes available. II Methods II.1 Momentum-Resolved EELS calculations Our q-EELS simulations are performed in a manner similar to that of Zeiger Zeiger2021 and Castellanos-Reyes CR2025 . In either case, molecular dynamics (MD) simulations are used to acquire time-dependent atomic configurations, which are then used as the input for frozen phonon multislice (electron wave propagation) simulations. While the electron wave simulations are elastic (no energy loss is simulated), the result contains information on the frequency-dependent vibrations from MD and the scattering probabilities of the transmitted electrons Forbes2010 . Two slightly differing methods for these simulations have been developed, the first of which was by Zeiger et al. Zeiger2021 . Given the frequency-resolved nature of these simulations and the use of multislice simulations for the electron wave, they have been referred to as “Frequency-Resolved Frozen-Phonon Multislice” or “FRFPMS”. These simulations originally made use of a custom frequency-specific thermostat to generate atomic configuration snapshots corresponding to a given frequency, however subsequent works simply use a band-pass filter over a single MD simulation. The resulting frequency-filtered atomic configurations are then used for frozen phonon multislice simulations. The q-EELS signal for a given energy bin is then obtained by taking the difference between the incoherent and coherent sum across multiple frozen phonon configurations. In the work of Castellanos-Reyes et al. CR2025 , multislice simulations are performed over consecutive timesteps, and a Fourier transform over time yields the energy-resolved q-EELS signal. The Fourier transform over the multislice exit wave is referred to as the “time autocorrelation of the auxiliary wave” and this approach has thus been referred to as the “TACAW” method. TACAW yields similar results to FRFPMS, but it is much more computationally efficient. FRFPMS requires multiple frozen phonon configurations for each energy (where parameters such as convergence-angle and sample thickness affect the number of configurations required). By comparison, the number of time-steps used in TACAW directly translates to the frequency resolution. In this work, we use the TACAW method (the post-hoc Fourier transform approach developed by Castellanos-Reyes et al. CR2025 ). We also performed simulations using FRFPMS (the frequency-binning method developed by Zeiger et al. Zeiger2021 ) to ensure observations were maintained (an example is available in the Supplemental Material). We use the Large-scale Atomic/Molecular Massively Parallel Simulator software (LAMMPS LAMMPS ) for our molecular dynamics (MD) simulations, and abTEM abTEM for our multislice simulations. None of our findings should be unique to the simulation method (“FRFPMS” vs. “TACAW”), or the software packages used (LAMMPS, or abTEM). We will thus refer to the combined molecular dynamics and frequency-resolved elastic multislice simulations as “q-EELS simulations”, as the precise method or software should not affect the overall results. Our q-EELS simulations begin with molecular dynamics (MD), with the velocities and displacements exported at frequent intervals (chosen based on the maximum phonon frequency expected within the system). All of our MD is performed under the NVE ensemble at room temperature with periodic boundary conditions. Periodic boundary conditions do not imply that an infinite volume is simulated however, as the length of the system constrains what modes are present. For a crystal with N atoms along a given direction, there are N modes along each branch within the first Brillouin zone. Simulation sizes are therefore selected based on a balance of reciprocal space resolution (favoring larger simulations) and computational abilities (limiting the maximum volume being simulated). Details on the exact simulations will be listed in subsequent sections. The suitability of the periodic boundaries in the through-plane direction will also be discussed. Multislice simulations are then performed over the atomic configuration for each timestep. The multislice exit waves (containing all elastic electron scattering information) are recorded in reciprocal space. Exit waves can be recorded as a function of depth through the sample, meaning the result of the simulation is a 3D or 4D data cube with axes of \omega , k_{x} , k_{y} and optionally z . Post-processing of this

Read the complete text