A Finite Chunk of Topological Insulator Has a Measurable Edge Signal That Never Goes Away — and Now There's a Floor on How Big It Can Be

Proves that chiral edge modes of finite Chern-insulator flakes carry a minimum dipole weight ≥|C| per unit area, detectable by a coupled circuit depending on probe frequency.

Imagine you're trying to weigh a picture frame by putting it on a kitchen scale. If the frame is big enough, most of the weight comes from the wood — the bulk. But the glass edge-trim also weighs something, and no matter how you scale the frame up, that trim never becomes negligible relative to the frame's footprint. This paper is about that trim: the electromagnetic signature of the edge of a topological insulator, and specifically why it can never be zero. The committed claim: a finite flake of a Chern insulator — a 2D topological material with chiral edge states — carries an 'edge quantum weight' K_edge that satisfies K_edge ≥ |C| per unit area (where C is the Chern number), and this floor is an exact consequence of the positive absorption of both circular polarizations by each edge transition plus the known cancellation of circular dichroism between edge and bulk. This is not an approximation or a numerical observation. It is an analytical bound derived from first principles of linear response theory. The ladder context matters. The bulk quantum weight K_bulk is already known to be bounded below by |C| — that's the classic result linking the Chern number to optical spectral weight. What Narula shows is that the edge adds its own independent term with the same floor. A weakly coupled circuit operating slower than the edge modes reads K_bulk + K_edge; a faster circuit below the bulk gap reads only K_bulk. The order of limits — grow the flake first or slow the circuit first — determines what you measure. This is a genuinely new observation about the interplay of system size and probe timescale in topological systems. Architecturally, the proof lives in the free-fermion tight-binding world. The analytical argument chains together the f-sum rule, the positivity of both circular-polarization absorption channels, and the known result that circular dichroism of edge and bulk cancel. Exact numerical verification is done on lattice flakes (square and rectangular geometries) with free-fermion diagonalization. The edge weight per channel is a geometry-dependent number that the lattice calculations reproduce to within 0.003 — a remarkable level of agreement between the analytical prediction and exact numerics. Integrity here is strong for a theoretical condensed-matter paper. The proof is analytical, the numerics are exact (not approximate or variational), and the code and data are deposited on Zenodo with a DOI. The benchmarks are self-consistent: the analytical bound is checked against exact free-fermion calculations, not against an independent experiment. This is standard for the subfield — you prove a theorem, then verify it numerically on exactly solvable models. There is no cherry-picking because the bound is supposed to hold universally; the lattice calculations are verification, not the main result. The milestone question is about experimental detection. The paper explicitly notes that weak coupling is required, and a growing flake at fixed coupling eventually leaves the weak-coupling regime. The practical question is whether a real circuit-QED or microwave-cavity experiment can resolve K_edge on a finite Chern-insulator flake, and the paper gives you the frequency-dependent protocol to do it. The gap between 'exact free-fermion lattice' and 'real material with interactions, disorder, and finite temperature' is the chasm every topological-matter theorist faces. The obvious next experiment not run: interacting systems. The entire proof hinges on free-fermion structure. The authors acknowledge the weak-coupling limitation but do not attempt even mean-field or DMRG checks on interacting Chern insulators. My read is (a) — the interacting case is a genuinely harder problem requiring different machinery, not something you bolt on. A second gap: no disorder. Real edges are disordered; whether the floor survives Anderson-type disorder on the boundary is a natural and important follow-up.