A T-matrix scattering formalism for electron-beam spectroscopy

This paper presents a versatile T-matrix-based scattering formalism implemented in the open-source `treams_ebeam` software to efficiently simulate cathodoluminescence and electron energy-loss spectroscopy for complex nanophotonic structures interacting with fast electrons.

The Gist

Imagine you have a super-powerful flashlight, but instead of light, it shoots tiny, invisible particles called electrons. You want to use this "electron flashlight" to peek inside a tiny, complex world made of nanomaterials (things smaller than a hair) to see how they react to light. This is what scientists do in electron-beam spectroscopy.

However, there's a problem: The math required to predict how these electrons bounce off and interact with these tiny structures is incredibly difficult, like trying to solve a giant 3D puzzle where every piece changes shape as you touch it. Traditional computer methods are slow and get stuck easily, especially when you have many pieces (scatterers) arranged in patterns.

This paper introduces a new, clever tool called the T-matrix method to solve this puzzle faster and more accurately. Here is how it works, explained with simple analogies:

1. The Problem: The "Messy Room" vs. The "Organized Library"

Imagine you are trying to understand how sound waves bounce around a room full of furniture (the nanomaterial).

• Old Methods (FDTD, FEM): These are like trying to map every single air molecule in the room, one by one, as the sound wave hits every chair and table. It's incredibly detailed but takes forever to calculate. If you add more furniture, the time it takes explodes.
• The New Method (T-matrix): This is like having a catalog card for every piece of furniture. You don't need to re-calculate how a chair reacts to sound every time. You just look up its "reaction card" (the T-matrix), which tells you exactly how it scatters sound based on its shape and material. Once you have the card, you can reuse it instantly.

2. The Innovation: Teaching the "Electron Flashlight" to Speak the Language

The authors realized that while the T-matrix method is great for light, it wasn't set up to talk to fast-moving electrons.

• The Mismatch: Light usually travels in flat waves (like ripples on a pond), but a fast electron moving in a straight line creates a very specific, curved "cylindrical" wave pattern (like a ripple spreading out from a stick dropped in water).
• The Fix: The authors wrote a "translator" code. They figured out how to take the electron's unique cylindrical wave and translate it into the language the T-matrix catalog understands. Now, the computer can instantly say, "Ah, an electron is coming! Here is exactly how this sphere, this cylinder, or this cluster of disks will react."

3. The Toolkit: treams_ebeam

The authors didn't just write the theory; they built a free software tool called treams_ebeam (part of a larger suite called treams).

• Think of this software as a virtual laboratory.
• You can tell the computer: "Here is an electron beam moving at 70% the speed of light. Here is a cluster of 15 silicon disks arranged in a line."
• The software instantly calculates two things:
  1. Cathodoluminescence (CL): How much light does the object glow when hit? (Like seeing a firefly light up).
  2. Electron Energy Loss (EELS): How much energy does the electron lose? (Like seeing how much the electron slows down after hitting the object).

4. What They Discovered (The Examples)

They tested their tool on three scenarios to prove it works:

• The Solo Act: They hit a single sphere, a wire, and a weirdly shaped disk. The tool perfectly predicted the "notes" (colors/energies) the objects would sing when hit by the electron.
• The Choir (Periodic Chain): They arranged many disks in a line. When hit by an electron, the disks didn't just act alone; they started "singing together" (resonating). The tool showed that as you add more disks, the sound becomes sharper and more directional, creating a specific type of light called Smith-Purcell radiation (think of it like a sonic boom of light).
• The Crowd (2D Grid): They arranged spheres in a square grid. They found that the electron only really "talks" to the spheres right next to its path. The ones far away barely react because the electron's influence fades quickly (like a whisper that doesn't carry far).

Why Does This Matter?

Before this, designing new nanodevices (like super-efficient solar cells or tiny lasers) required running slow, heavy simulations that could take days.

• Speed: This new method is much faster because it reuses the "catalog cards" (T-matrices).
• Versatility: It works for single objects, messy clusters, and perfect grids.
• Design: It allows engineers to quickly test thousands of designs on a computer before building them in a lab.

In a nutshell: This paper gives scientists a fast, reusable "cheat sheet" to predict how fast electrons interact with tiny, complex materials. It turns a slow, tedious math problem into a quick, accurate calculation, unlocking the ability to design the next generation of nanotechnology light sources.

Technical Summary

1. Problem Statement

Electron-beam spectroscopy, specifically Cathodoluminescence (CL) and Electron Energy-Loss Spectroscopy (EELS), has become indispensable for probing the optical properties of nanophotonic materials with high spatial and temporal resolution. Unlike traditional light-based methods, fast electrons interact with the near-field of photonic structures via evanescent fields, allowing them to excite both radiative and "dark" (non-radiative) optical modes across a broad spectrum (from plasmons to phonons).

However, as photonic platforms grow in complexity (e.g., periodic arrays, metasurfaces, and clusters of arbitrary shapes), existing numerical simulation tools face significant limitations:

• Computational Cost: Methods like Finite-Difference Time-Domain (FDTD), Finite Element Method (FEM), and Boundary Element Method (BEM) require solving Maxwell's equations in discretized space/time. This becomes prohibitively expensive for systems with multiple length scales or when extensive parameter sweeps are needed.
• Periodicity Challenges: Modeling infinitely extended periodic structures is difficult for many standard discretization methods.
• Lack of Unified Framework: There is a need for a versatile, algebraic framework that can efficiently handle interactions between fast electrons and complex assemblies of scatterers (both periodic and aperiodic) without re-computing the entire system for every illumination condition.

2. Methodology

The authors propose a general framework based on the T-matrix method (Transition matrix method) extended to describe interactions with fast electrons. The core theoretical and implementation steps include:

A. Theoretical Framework

1. Field Representation:

   • The electromagnetic field of a relativistic electron moving along a straight trajectory is expressed in the Cylindrical Wave Basis (CWB). The field is identified as a purely Transverse Magnetic (TM) cylindrical wave of singular type (order $m=0$) with a fixed longitudinal wavevector $k_z = \omega/v$.
   • To interact with scatterers, this singular cylindrical wave is converted into regular cylindrical waves and subsequently expanded into the Spherical Wave Basis (SWB) using connection formulas (translation coefficients). This allows the electron field to be treated as an incident field within the standard T-matrix formalism.

2. Scattering Formalism:

   • Single Scatterers: The interaction is governed by the relation $\mathbf{p} = \mathbf{T}\mathbf{a}$, where $\mathbf{a}$ are incident expansion coefficients, $\mathbf{p}$ are scattered coefficients, and $\mathbf{T}$ is the T-matrix encoding the object's geometry and material properties.
   • Multiple Scatterers (Clusters): For ensembles of $N$ scatterers, the method uses translation coefficients to account for multiple scattering. The local T-matrix of the cluster is derived algebraically: $\mathbf{T}_{local} = [\mathbf{I} - \mathbf{T}_{diag}\mathbf{C}^{(3)}]^{-1}\mathbf{T}_{diag}$.
   • Periodic Structures: For 1D periodic chains, the formalism incorporates phase factors ($e^{ik_z\Lambda}$) to sum interactions over infinite lattice sites, allowing for the calculation of collective resonances (e.g., lattice modes, bound states in the continuum).

3. Observable Calculations:

   • CL Probability ($\Gamma_{CL}$): Calculated by integrating the Poynting vector of the scattered field over a surface enclosing the structure (sphere for SWB, cylinder for CWB).
   • EEL Probability ($\Gamma_{EEL}$): Calculated by integrating the work done by the scattered field on the electron along its trajectory.
   • The framework provides unified expressions for these probabilities in both CWB (per unit length) and SWB (per particle).

B. Software Implementation

• The formalism is implemented as an open-source Python module, treams_ebeam, extending the existing treams library.
• Key Features:
  • Initialization of electron beams with specific wavenumber, velocity, and impact parameter.
  • Automated basis conversion (CWB to SWB) for incident fields.
  • Functions to compute CL and EEL probabilities directly from T-matrices.
  • Support for arbitrary scatterer shapes (via pre-computed T-matrices from other solvers like BEM) and periodic arrangements using the Ewald summation method.

3. Key Contributions

1. Generalized T-Matrix Formalism for Electrons: The paper establishes a rigorous mathematical link between the cylindrical wave representation of fast electrons and the T-matrix scattering theory, enabling the simulation of electron-beam spectroscopy for arbitrary scatterer geometries.
2. Efficient Handling of Periodicity: The method efficiently models infinite periodic chains and 2D/3D arrays by leveraging algebraic formulations and Ewald summation, avoiding the computational bottlenecks of supercell approaches in FDTD/FEM.
3. Open-Source Tool (treams_ebeam): The authors release a user-friendly, open-source codebase that integrates electron-beam physics into a mature scattering library, lowering the barrier to entry for simulating CL and EELS.
4. Unified CL/EELS Prediction: The framework simultaneously predicts radiative (CL) and non-radiative (EELS) channels, facilitating the study of dark modes and energy dissipation mechanisms.

4. Results and Demonstrations

The authors validate the framework through three representative examples:

• Single Scatterers:

  • Dielectric Sphere: Successfully reproduced magnetic/electric dipole and quadrupole modes. The difference between EEL and CL curves correctly quantified absorption losses.
  • Metallic Nanowire: Simulated Surface Plasmon Polaritons (SPPs) as Lorentzian peaks in the spectrum, confirming the excitation of TM cylindrical waves.
  • Amorphous Silicon Nanodisk: Demonstrated the ability to handle non-spherical geometries by importing T-matrices from BEM solvers, revealing mixed multipolar modes.

• Periodic Chain of Nanodisks (1D):

  • Simulated a chain of elliptical nanodisks parallel to the electron beam.
  • Lattice Resonance: Observed the emergence of a sharp collective resonance at ~1.7 eV as the number of particles increased.
  • Smith-Purcell Radiation: In the infinite chain limit, the CL emission became highly directional, matching the theoretical prediction for Smith-Purcell radiation, while the EEL spectrum showed the suppression of radiative decay (dark mode behavior).

• Finite 2D Array of Nanospheres:

  • Modeled a cluster of aluminum nanospheres.
  • Mode Hybridization: Observed the splitting of single-particle plasmon modes into bonding and antibonding hybridized modes due to near-field coupling.
  • Evanescent Coupling: Demonstrated that the electron beam primarily excites nanoparticles close to its trajectory due to the evanescent nature of the field, limiting the effective size of the excited lattice resonance.

5. Significance

This work represents a significant advancement in the theoretical modeling of electron-matter interactions:

• Design & Engineering: It provides a fast, accurate tool for designing next-generation nanophotonic devices (e.g., nanoantennas, metasurfaces) where electron-beam spectroscopy is used for characterization.
• Physical Insight: By unifying fast-electron physics with advanced scattering theory, the framework allows researchers to disentangle radiative and non-radiative decay pathways, crucial for understanding bound states in the continuum (BICs) and other complex collective phenomena.
• Scalability: The T-matrix approach offers a scalable alternative to grid-based methods, particularly for systems involving multiple length scales or requiring extensive parameter optimization, making it ideal for the inverse design of nanophotonic structures.

In summary, the paper bridges the gap between theoretical scattering formalism and practical electron-beam spectroscopy, offering a versatile computational engine for the nanophotonics community.