Multiple-Scattering Framework

• Multiple-scattering frameworks are computational methodologies that model repeated wave interactions in heterogeneous media, capturing coherent interference and energy transfer.
• They utilize coupled-channel, T-matrix, and multislice methods to solve wave equations and account for complex phenomena like Bragg diffraction and localization.
• These frameworks are essential for advanced imaging, tomographic reconstructions, and resonant structure analysis across optics, electron microscopy, and nanophotonics.

A multiple-scattering framework is a set of theoretical and computational methodologies for modeling the physical process in which waves, particles, or fields are scattered more than once within a medium containing inhomogeneities, interfaces, or structured components. Such frameworks are essential when the single-scattering (or kinematical) approximation fails—specifically, when the strength, density, or geometry of scatterers leads to constructive or destructive interference between multiply scattered paths. These frameworks are crucial across domains, including condensed matter physics, optics, materials science, electron microscopy, photonics, and beyond, as they enable accurate prediction, inversion, and interpretation of observables in strongly or coherently scattering systems.

1. Core Mathematical Structures and Problem Classes

Multiple-scattering frameworks are based on solutions to linear or nonlinear wave equations, typically the scalar or vector Helmholtz equation, or corresponding time-domain (elastodynamic, Dirac, or radiative transfer) equations. The general form

$(\nabla^2 + k^2 n^2(\mathbf r))\psi(\mathbf r) = 0$

governs electromagnetic, acoustic, or matter-wave propagation in spatially varying media. The frameworks differ primarily by the system geometry, excitation, and the nature of the scatterers (atoms, grains, inclusions, boundaries). Multiple scattering manifests via the breakdown of single-pass models and requires formalisms that capture the coherent buildup of scattered waves, energy transfer between different orders, and their dynamic or statistical interaction.

Key problem classes include:

• Periodic or quasi-periodic multi-scatterer coupling (photonic/metamaterial lattices)
• Wave transport in disordered/heterogeneous media (localization, optical diffusion, backscattering)
• Bragg diffraction in crystals (electron, x-ray, neutron)
• Inverse-scattering and imaging in strongly scattering volumes (biological tissue, composites)
• Electron, phonon, or quasiparticle scattering in solids (density of states, transport)
• Radiative transfer and energy migration in participating media

2. Coupled-Channel and Transfer-Matrix Approaches

A central approach is the conversion of the governing equation into a finite system of coupled-mode or transfer equations, enabling explicit tracking of amplitude exchange between propagating eigenchannels. For Bragg diffraction in crystals, the Howie–Whelan (or N-beam dynamical diffraction) equations represent such a formulation:

$$\frac{d}{dz}\phi_g(z) = i s_g \phi_g(z) + i \sum_{g' \neq g} \frac{1}{2\xi_{g-g'}} \phi_{g'}(z)$$

where $\phi_g(z)$ are mode amplitudes, $s_g$ are excitation errors, and $\xi_{g-g'}$ are extinction distances encoding Fourier-coupling by the crystal potential. In matrix form,

$$\frac{d}{dz}\Phi(z) = i A \Phi(z), \qquad \Phi(z) = (..., \phi_{g_1}, \phi_{g_2}, ...)$$

with Hermitian $A$. Solutions employ diagonalization, leading to intensity computations $I_g(L) = |\phi_g(L)|^2$ after the wave traverses thickness $L$. This formalism is generalizable to all coherent, layered, or periodic structures where inter-channel coupling is finite or can be efficiently truncated. Dynamical diffraction theory strictly supersedes the kinematical (single-scattering) regime when sample thickness or structural coherence exceeds the relevant extinction length, generally for thick or high-quality crystals (Vallejo et al., 2017).

3. Foldy–Lax, T-matrix, and Green’s Function Resummation

For finite or infinite clusters of discrete scatterers, the Foldy–Lax multiple-scattering formalism provides a systematic expansion for the Green's function in terms of single-site $T$-matrices and mutual interaction tensors via the Dyson or Lippmann–Schwinger equations. For example, the total Green’s function with $N$ scatterers is

$$G = G_0 + G_0 \sum_i T^{(i)} G_0 + G_0 \sum_{i,j} T^{(i)} G_0 T^{(j)} G_0 + \cdots$$

This can be recast in matrix form as

$$\mathcal{T}_{\rm tot} = \mathcal{T} [1 - \mathcal{G}_0 \mathcal{T}]^{-1}$$

where $T^{(i)}$ encapsulate the local scattering properties, and $\mathcal{G}_0$ encodes free-space or host-medium propagation between centers. This resummation captures all orders—diffuse, recurrent, and coherent scattering—and is foundational in photonic- and electronic-structure calculations (Xu et al., 2016, Nečada et al., 2020, Vaishnav et al., 2010, Fischbach et al., 9 Sep 2024).

In periodic arrays, long-range coupling is efficiently handled via Ewald summation, decomposing the lattice propagator into rapidly converging real- and reciprocal-space sums (Nečada et al., 2020). T-matrix frameworks also enable extraction of system resonances, with contemporary adaptation empowering rapid and differentiable pole-finding via rational approximation (AAA algorithm) for gradient-based design (Fischbach et al., 9 Sep 2024).

4. Multislice, Born-Series, and Volume Integral Methods

Volume scattering in slab or extended geometries is addressed by stratifying the domain into thin slices (“multislice” approach), with each layer contributing local scattering and field propagation performed explicitly between slices. The multi-slice method for the scalar or vectorial field takes the generic form

$$U_{j+1}(x,y) = \mathcal{F}^{-1}\left\{ \mathcal{F}[U_j(x,y)] \cdot H(f_x,f_y;\Delta z) \right\}$$

$\text{modulated by } \chi_j(x,y)$

and is recursively applied, yielding the built-up field and, through the nonlinear dependence, all multiple-scattering orders. Accurate models account for polarization, anisotropy, and non-paraxial propagation (Rogalski et al., 1 Aug 2025, Zhu et al., 2022, Mu et al., 2022, Tahir et al., 2018).

High-throughput implementations exploit FFT-based algorithms and domain-decomposition, achieving gigavoxel-scale 3D reconstructions including strong multiple scattering (Rogalski et al., 1 Aug 2025). In contrast, single-scattering (first Born/Rytov) approximations replace the field at each slice by the unperturbed incident field, neglecting amplitude recycling and phase retardation from prior layers—a limitation directly addressed by multi-slice nonlinear solvers (Rogalski et al., 1 Aug 2025, Zhu et al., 2022, Tahir et al., 2018).

5. Diffusion, Transport, and Radiative-Transfer Methods

When many scattering events randomize phase, a transport-theoretic perspective is appropriate. The radiative transfer equation (RTE) is systematically expanded in spherical harmonics (“$P_N$ method”), producing a coupled hierarchy of angular-moment PDEs for the radiance field, of which the $P_1$ (diffusion) or flux-limited diffusion (FLD) equations are special cases (Koerner et al., 2018, Koerner et al., 2014):

$$L(\mathbf{x}, \Omega) \approx \sum_{l=0}^N \sum_{m=-l}^{l} \psi_l^m(\mathbf{x}) Y_l^m(\Omega)$$

Enforcing the causal bound $|E|\leq\phi$ introduces a nonlinear “flux limiter” which is necessary in optically thin or inhomogeneous media where standard diffusion breaks down (Koerner et al., 2014, Koerner et al., 2018). FLD and higher-order $P_N$ schemes balance computational tractability and accuracy, transitioning from deterministic diffusive models to full Monte Carlo or stochastic radiative methods when required.

6. Advanced Iterative and Domain-Decomposed Methods

For large, inhomogeneous, or multi-obstacle settings, iterative domain-decomposition approaches enclose each scatterer (or cluster) in artificial boundaries, solve local (single-scatterer) problems, and couple them through boundary traces and outgoing matching conditions. The resulting interface problems involve smooth (Fredholm-II) kernels and compact off-diagonal operators, solvable by robust Krylov subspace iterations with exponential convergence and high parallel scalability (Xie et al., 2019, Acosta, 2013). On-surface radiation condition (OSRC) frameworks further reduce the dimensionality of exterior problems and, although approximate, serve as efficient preconditioners for boundary-integral equations in the high-frequency or weak-coupling regime (Acosta, 2013).

7. Physical Effects, Regimes of Validity, and Inversion Applications

Multiple-scattering frameworks enable quantitative prediction of phenomena including:

• Bragg-peak intensification and non-Debye–Waller dynamics (thermal, non-equilibrium crystals) (Vallejo et al., 2017)
• Resonance formation and band-structure engineering (nanophotonic, excitonic, or plasmonic systems) (Fischbach et al., 9 Sep 2024, Nečada et al., 2020)
• Enhanced or suppressed attenuation and velocity dispersion in polycrystals (SOSA corrections) (Roy et al., 9 May 2025)
• Nontrivial coherent effects: coherent back-scattering, Anderson localization, pseudospin rotation and LDOS modulations (disordered optics, graphene) (Cherroret, 2018, Vaishnav et al., 2010)
• Deep, label-free tomographic imaging in strongly scattering biological tissues, leveraging full multiple-scattering inversion for artifact-free 3D reconstructions at gigavoxel scale (Rogalski et al., 1 Aug 2025)

Validity regimes are dictated by:

• Scattering strength, thickness, and structural correlation relative to mean-free-path, extinction, or correlation length
• Incident energy, coherence, and geometry (forward, back, or off-axis illumination)
• System disorder, uniformity, and boundary conditions

When multiple scattering dominates ($L\gtrsim 2\pi\xi_g$ or optical thicknesses $\gtrsim$1), dynamical frameworks are necessary; in the dilute or small-thickness limit, kinematical or first-Born approaches suffice.

Conversely, the sensitivity of multiple-scattering observables to dynamic or static alterations in scatterer properties (e.g., transient defect population, non-equilibrium strain, high-order phase modulation) enables inversion and design—such as time-resolved UED analysis, rational resonance design with automatic differentiation, or 3D refractive-index tomography using optimization-driven multiplicative updates.

---

References:

• "Observation of large multiple scattering effects in ultrafast electron diffraction on single crystal silicon" (Vallejo et al., 2017)
• "High-fidelity intensity diffraction tomography with a non-paraxial multiple-scattering model" (Zhu et al., 2022)
• "Gigavoxel-Scale Multiple-Scattering-Aware Lensless Holotomography" (Rogalski et al., 1 Aug 2025)
• "Multiple-scattering T-matrix simulations for nanophotonics: symmetries and periodic lattices" (Nečada et al., 2020)
• "On-surface radiation condition for multiple scattering of waves" (Acosta, 2013)
• "An Efficient Iterative Method for Solving Multiple Scattering in Locally Inhomogeneous Media" (Xie et al., 2019)
• "Flux-Limited Diffusion for Multiple Scattering in Participating Media" (Koerner et al., 2014)
• "Coherent multiple scattering of light in (2+1) dimensions" (Cherroret, 2018)
• "Holographic particle localization under multiple scattering" (Tahir et al., 2018)
• "A multislice computational model for birefringent scattering" (Mu et al., 2022)
• "Multiple Scattering of Elastic Waves in Polycrystals" (Roy et al., 9 May 2025)
• "A framework to compute resonances arising from multiple scattering" (Fischbach et al., 9 Sep 2024)
• "Intravalley Multiple Scattering of Quasiparticles in Graphene" (Vaishnav et al., 2010)
• "An improvement of the Moliere-Fano multiple scattering theory" (Tarasov et al., 2011)
• "$P_N$-Method for Multiple Scattering in Participating Media" (Koerner et al., 2018)
• "Compton Scattering in Plasma: Multiple Scattering Effects and Application to Laser-Plasma Acceleration" (Kumar et al., 2013)