Nonlocal Optical Response and Surface Susceptibilities: A Systematic Derivation via Spatial Moment Expansion
F. Zolla · 2026
Read Nonlocal Optical Response and Surface Susceptibilities: A Systematic Derivation via Spatial Moment Expansion by F. Zolla (2026) 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-2605-15716.txt.
Introduction
Frédéric Zolla links nonlocal kernels to surface susceptibilities for curved interfaces. From the general nonlocal relation, he combines a moment spread with a thin-layer limit. Interface detail shrinks into one scalar: the surface value, equal for both field components. It generalizes the Feibelman d-parameters to curved faces. Bend fixes, tied to bend H and bend K, are derived. Cases cover planar, round, tube, and oval shapes for Gaussian, Yukawa, and Lorentz kernels. Generalized Maxwell boundary rules follow, versus Fresnel results.
Catalog record
| CREATOR | F. Zolla |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 115440 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.15716
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Nonlocal Optical Response and Surface Susceptibilities: A Systematic Derivation via Spatial Moment Expansion F. Zolla Abstract We present a systematic theory connecting the nonlocal response kernel of a homogeneous medium to its effective surface susceptibilities for an arbitrary curved interface. Starting from the most general tensorial nonlocal constitutive relation and combining a spatial moment expansion with a distributional thin-layer limit, we show that the full complexity of the interfacial response condenses, at leading order, into a single scalar: the surface susceptibility χ s \chi^{s} , equal for the tangential and normal components of the electric field. These quantities provide a constructive generalization of the Feibelman d d -parameters to interfaces of arbitrary curvature, and the curvature corrections, proportional to the geometric invariants H H (mean curvature) and K K (Gaussian curvature), are derived explicitly. The formalism is illustrated on a comprehensive set of analytically tractable cases (planar, spherical, cylindrical, and ellipsoidal interfaces) for several kernel choices (Gaussian, Yukawa, tensorial Lorentz). Generalized Maxwell boundary conditions are established and compared with the classical Fresnel results. 1 Introduction The interaction between light and matter is one of the oldest subjects in physics. Despite the abundance of work devoted to this question over more than a century, it would be an overstatement to claim that the field is fully understood today. In many practical situations, theoretical descriptions still rely on partially empirical models whose microscopic justification remains incomplete. One of the deep reasons for this state of affairs lies in the coexistence of several vastly different spatial scales in optical phenomena. In a typical problem of electromagnetism applied to optics, one can distinguish at least three characteristic scales. • The first is the atomic scale, of the order of the angström. At this scale, electromagnetic fields vary over distances comparable to the dimensions of electronic orbitals. Quantum mechanics reigns supreme. • The second is the scale of optical wavelengths, typically of the order of a micrometre in the visible range. • The third is the scale of the macroscopic objects being illuminated, which can reach several millimetres or centimetres. A complete description of the light–matter interaction should in principle allow these three levels of description to be connected. In particular, the passage from the atomic scale to the optical scale should be obtained through a procedure of statistical and quantum averaging, making it possible to define the macroscopic quantities used by the optical physicist, such as the permittivity or the susceptibility of the medium [ 1 , 2 ] . In practice, this programme remains largely unfulfilled for the systems of interest to modern optics. Ab initio approaches exist, but they remain difficult to apply in realistic situations involving surfaces, interfaces, or complex structures [ 3 ] . One is therefore led to adopt a different approach, often described as phenomenological, which consists in postulating certain constitutive relations between the macroscopic electromagnetic quantities. More precisely, one seeks to establish relations connecting the electric and magnetic fields (\mathbf{E},\mathbf{B}) to the displacement and induction fields (\mathbf{D},\mathbf{H}) . These relations must satisfy several fundamental requirements. In particular, they must: • be independent of the choice of coordinate system; • respect the symmetries of the physical system under consideration; • be compatible with the general principles of physics, in particular the causality principle. In this framework, the electric polarisation \mathbf{P} plays a central role, since it connects the fields \mathbf{E} and \mathbf{D} through the relation \mathbf{D}=\varepsilon_{0}\mathbf{E}+\mathbf{P}. The main difficulty then consists in determining the functional relation between \mathbf{P} and \mathbf{E} . The optical response of material media is traditionally formulated within the local framework, in which the polarization vector at a point \mathbf{r} depends only on the electric field at that same point [ 1 , 2 ] . This approximation, which is remarkably successful for describing wave propagation in macroscopic homogeneous media, breaks down as soon as the relevant length scales become comparable to the microscopic characteristic lengths of the material, typically a few Ångströms to a few nanometres, for metals and semiconductors. Spatial non-locality is a considerably more subtle issue. In this case, the polarisation vector at a point in space depends on the electric field in a spatial neighbourhood. Unlike the temporal case, there is no principle of spatial causality that could constrain the form of this relation. Furthermore, the three-dimensional structure of space allows for a wide variety of geometric behaviours that have no counterpart in the temporal problem. Spatial dispersion which translates the dependence of the dielectric susceptibility not only on frequency \omega but also on the wavevector \mathbf{k} was introduced theoretically by Pekar [ 4 ] and developed systematically by Agranovich and Ginzburg [ 5 ] . It expresses the fact that the polarization at \mathbf{r} results from a weighted integral of the field over a neighbourhood of \mathbf{r} , whose spatial extent is set by the range of the response kernel. This nonlocal framework has proved indispensable for describing excitonic polaritons in semiconductors [ 6 , 7 ] , and more recently, nonlocal effects in plasmonic nanostructures [ 8 , 9 , 10 ] . The rapid growth of nanophotonics has renewed interest in the description of interfaces in this nonlocal context. Even in a centrosymmetric medium, the symmetry breaking induced by a surface generates a response concentrated in an atomic-scale transition layer, qualitatively different from the bulk. Feibelman [ 11 ] showed that this surface response can be characterized, at leading order, by two scalar parameters d_{\perp} and d_{\parallel} , the so called Feibelman d -parameters , encoding the response of the transition layer to the normal and tangential components of the electric field. These parameters have since been computed ab initio for a wide range of metal surfaces [ 12 , 13 ] , measured experimentally via ellipsometry and differential reflectance spectroscopy [ 14 , 15 ] , and incorporated into extended Maxwell boundary conditions for various geometries [ 16 , 17 ] . Despite this progress, the connection between the bulk nonlocal response kernel and the effective surface susceptibilities, and more generally the generalized boundary conditions for an arbitrarily curved interface, has remained elusive. Most existing approaches rely either on phenomenological models such as the nonlocal hydrodynamic model [ 7 , 8 , 18 ] , or on material-specific microscopic calculations [ 12 , 19 ] , without providing a constructive link between the bulk kernel and the surface response for an arbitrary geometry. In particular, the influence of interface curvature which is central for understanding the optical properties of nanoparticles [ 10 , 16 ] and the nonlocal Mie corrections has not been derived systematically from the bulk kernel alone. The present paper fills this gap by providing a systematic and rigorous derivation of the linear surface susceptibilities from the bulk nonlocal response kernel. Our approach rests on two ingredients: 1. A spatial moment expansion of the tensorial kernel \Delta^{(1)}_{ij}(\mathbf{r},\mathbf{r}^{\prime}) , which extends the classical Taylor expansion of the field to the three-dimensional vector case and naturally decomposes the response into a bulk contribution and a boundary contribution; 2. A distributional thin-layer limit , which converts functions concentrated near the interface into singular distributions supported on \partial\Omega , i.e. Dirac deltas and their normal derivatives, whose coefficients