REVIEW 3 major objections 3 minor 2 cited by
Observation of flat-band skin effect
T0 review · 3 major / 3 minor · reviewed 2026-08-03 · deepseek-v4-flash
Pith's one-line read An ideal flat band can host a non-Hermitian skin effect, one that appears only when the surrounding dispersive bands enclose it in a point gap and disappears again at large non-Hermiticity.
desk verdict The flat-band skin effect is a real and well-argued theoretical result; the experiment confirms it but is not fully independent because the same response data were used to extract the model parameters. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The Green's function G(E) = (E - H_OBC)^{-1} evaluated infinitesimally close to the flat-band energy, together with the Lehmann (biorthogonal) spectral representation. The flat band is enforced by a rank-2 Hamiltonian structure, giving an exactly degenerate zero-energy subspace spanned by compact localized states; the Green's function response is determined by the left eigenvectors, which in the FBSE regime are exponentially localized at the boundary opposite to where the right (CLS) modes sit. The point-gap topology of the dispersive bands, not of the flat band, selects the parameter region where this happens.
What would settle it
Measure the flat-band steady-state response in the original (unrotated) lattice for parameters in region III while suppressing the dispersive gain modes (e.g., by using a narrow probe or time-gating); if the response still localizes at the edge, the FBSE does not disappear, contradicting the central re-entrant claim. Alternatively, sweep gamma_2 across the predicted boundary in Fig. 1(b) and check that the center-of-mass chi jumps sharply at the PBC gap-closing curve; a shift or absence of the jump would falsify the point-gap mechanism.
Extended reading notes
Core claim
The central claim is that a flat-band skin effect (FBSE) occurs when the flat band lies inside the point gap of the dispersive bands under periodic boundary conditions. Although the flat band's own open-boundary eigenvectors (compact localized states) remain spread across the bulk, a source at the flat-band frequency produces a response exponentially localized at one edge because some left eigenvectors become localized at the opposite boundary with very large magnitudes. The mechanism is biorthogonal: the Green's function at zero energy is a sum over flat-band modes, and the extreme non-normality of the flat-band subspace, rather than any winding of the flat band itself, produces the skin re
Load-bearing premise
The whole experiment assumes the measured steady-state response is produced by the ideal tight-binding model with couplings fitted from the same responses, and that the spectral-rotation trick used at large non-Hermiticity does not change the response physics; if either fails, the claimed observation is not established.
Editorial extensions
If this is right
- FBSE should be generic across flat-band models: the authors show it in AB cage, ladder, and Lieb lattices as well.
- The effect is re-entrant: it can be switched off by increasing non-Hermiticity, unlike standard skin effects.
- The gap-closing points are order-3 exceptional points under both periodic and open boundary conditions, and the flat-band quantum distance is discontinuous there.
- An active mechanical lattice reproduces the predicted responses, including the disappearance in region III when viewed through a spectrally rotated model.
- The response center-of-mass chi provides a basis-invariant diagnostic for FBSE, avoiding ambiguities from the degenerate flat-band subspace.
Reading between the lines
- If FBSE is tied to the point-gap enclosure, similar effects should appear in higher-dimensional flat-band systems where the dispersive bands form point-gap loops; a 2D test would be a natural next step.
- The biorthogonal mechanism implies the skin response is sensitive to the source's overlap with the localized left eigenvectors; pumping schemes that excite different flat-band superpositions could control the edge on demand.
- The 'disappearance at large non-Hermiticity' might be masked in the original model by dispersive gain modes with near-zero real frequency; the paper's spectral-rotation argument needs an independent check that isolates the flat band in the unrotated lattice.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper introduces and experimentally claims the flat-band skin effect (FBSE) in a one-dimensional non-Hermitian lattice. The model is a rank-defective three-band Hamiltonian with an exactly flat band at E=0 and two dispersive bands whose PBC spectra form loops in the complex-energy plane. The central theoretical claim is that FBSE appears exactly when the flat band lies inside the point gap of the dispersive bands, and disappears when the line gaps reopen at large non-Hermiticity. The theoretical analysis uses Green's functions, left/right eigenvector decomposition, the generalized Brillouin zone, and identifies order-3 exceptional points on the non-Bloch wavevector plane. The experimental part reports steady-state mechanical-lattice responses for three parameter regimes, claiming edge-localized response in the FBSE regime and its absence outside, including in the large-non-Hermiticity regime using a unitarily rotated model.
Significance. If the central claim is correct, it identifies a genuinely new mechanism for non-Hermitian skin effects: a topologically trivial flat band acquires skin-like response through the point-gap topology of surrounding dispersive bands, with a tunable, re-entrant parameter window. The theoretical framework is largely self-contained and parameter-free: the flat band is exact by construction, the FBSE criterion is derived from PBC spectra, and the CLS decomposition is given explicitly. The extension to AB-cage, ladder, and Lieb models in the Supplemental Material strengthens the generality of the criterion. The experimental platform is appropriate and capable of realizing the required non-reciprocal couplings. However, the experimental validation as reported is not independent: parameters are extracted from the same response data used for comparison, no uncertainties are given, and the key disappearance regime is measured in a rotated model rather than in the original OBC model. These issues are load-bearing for the experimental 'observation' claim.
major comments (3)
- [Fig. 3 caption and Table 1] The caption states that 'all effective parameters were extracted by fitting experimental responses to the Green’s function.' Therefore the excellent agreement between the measured flat-band responses and the Green’s-function calculations in Fig. 3(c1-c3) is partly tautological: the same data determine both the model parameters and the comparison. Table 1 lists retrieved t1, t2, γ1, γ2 without uncertainties. To establish the FBSE observation convincingly, the parameters should be calibrated by an independent measurement (e.g., two-site transmission, band-structure tracking, or a source configuration not used in the fit) and then used to predict the response out of sample. Without this, the experimental confirmation of the central phase diagram is not independent.
- [End Matter, Eqs. (5)-(6) and Fig. 3(b3)] The disappearance of FBSE in region III is experimentally demonstrated in the rotated Hamiltonian H3, obtained from H_OBC by H1 = i H_OBC and a subsequent unitary embedding, not in the original OBC model Eq. (2). While H3 is related by exact unitary transformations, a local single-site drive in H3 corresponds, in the original lattice, to a nonlocal superposition of sources. Thus the measured absence of edge response in H3 does not automatically establish the absence of FBSE for a local drive in H_OBC. The theoretical Fig. 1(b) already predicts the disappearance, but the experimental support is one step removed. The authors should specify the source mapping and show that the relevant response observable is preserved, or measure the original model with a scheme that isolates the flat band from the gain modes.
- [Fig. 3 and Table 1] No raw data, noise floor, or repeated-measurement statistics are reported. The experimental profiles in Fig. 3(b1-b3) are single normalized traces without error bars, and Table 1 gives no uncertainties in the fitted parameters. Since the mechanical lattice is active and feedback-controlled, systematic uncertainties in γ1 and γ2 are especially relevant for locating the regimes in Fig. 1(b). At minimum, the authors should report measurement uncertainties and demonstrate that the three regimes are separated by more than the experimental resolution in parameter space.
minor comments (3)
- [Eq. (3)] The definition of χ as written omits the normalization denominator. It should be χ = Σ_n n |R_n|^2 / Σ_n |R_n|^2 (or equivalently state that |R⟩ is normalized). As printed, the quantity is not a center-of-mass measure.
- [Fig. 2 caption] The caption describes the REVs as 'CLS' in panels (a,c), while the main text states that the orthogonalized REVs are 'still not exactly CLS.' The caption should be reconciled with the main text to avoid confusion.
- [References [54] and [69]] References [54] and [69] appear to be explanatory footnotes embedded in the reference list rather than standard citations. These should be moved into the main text or footnotes and formatted consistently.
Circularity Check
Theoretical FBSE derivation is self-contained; only the experimental Green's-function comparison is partly tautological because couplings were fitted to the same responses.
-
fitted input called prediction
[Experimental observation of the FBSE, Fig. 3 caption; End Matter, Table 1]
"In all experiments, t1 = −1.06, t2 = −0.3 and all effective parameters were extracted by fitting experimental responses to the Green's function."
The same measured steady-state responses were used to extract the effective couplings (Table 1) and then compared with Green's-function calculations using those couplings (Fig. 3(c1-c3), 'showing excellent agreement'). The agreement is therefore at least partly enforced by the fitting procedure rather than constituting an independent test of the FBSE. The theoretical prediction itself (Fig. 1) does not rely on these fitted values, and the raw displacement profiles (Fig. 3(b1-b3)) provide nontrivial qualitative evidence, so this is a partial, experimental circularity rather than a circular derivation.
full rationale
The central theoretical claim is not circular. Eq. (1) defines the model; the flat band follows directly from rank(H)=2, and the dispersive bands are solved explicitly as E(k)=±sqrt(Δ − γ1^2 − γ2^2 − 2iγ1 t2 sin k), with gap-closing positions obtained by setting these expressions to zero. The FBSE criterion (flat band inside the PBC point gap of the dispersive bands) is then checked by computing the OBC Green's function and the GBZ spectra from the same Hamiltonian, without fitted constants. No load-bearing self-citation chain is used: prior papers on the active mechanical platform are cited for the experimental apparatus, but the relevant setup is also described in the End Matter, and the unitary transformation to H3 is derived explicitly with matrices U1 and U2 rather than imported as an ansatz. The only genuine circularity is in the experimental validation loop: Fig. 3 states that all effective parameters were extracted by fitting experimental responses to the Green's function, and the same Green's function is used to generate the 'excellent agreement' curves. This makes the quantitative comparison semi-tautological, though the raw measured spatial profiles remain meaningful evidence. Because the theory itself is self-contained and the fitting loop does not enter the derivation of the FBSE condition, the overall circularity score is low.
Assumptions & free parameters
free parameters (2)
- Experimental t1, t2 =
-1.06 s^-1, -0.3 s^-1
- Experimental γ1,γ2 sets =
(0.62,0.32), (0.9,0.32), (1.5,0.66) s^-1
assumptions (5)
- domain assumption Biorthogonal completeness and Lehmann representation G(E)=Σ Eη^{-1}|ψ^R><ψ^L| for the non-Hermitian OBC Hamiltonian (ref [56]).
- domain assumption Generalized Brillouin zone method: OBC eigenstates are obtained from H(β) with β∈C and |β|>1 (refs [43,63]).
- domain assumption Point-gap topology governs ordinary NHSE in dispersive bands (refs [43,58]), so a single point (flat band) has trivial point-gap topology.
- domain assumption With OBC gapped at E=0, a source with energy Eη=0+i10^-20 excites only flat-band modes.
- domain assumption The active mechanical feedback loop implements the ideal non-reciprocal couplings with negligible latency and linear response (End Matter: latency 0.35 ms << 92 ms period).
Cite this review
Pith. "Pith review of Observation of flat-band skin effect." pith.science (2026). https://pith.science/paper/GJREFQ33
@misc{pith2026251219745,
author = {Pith},
title = {Pith review of: Observation of flat-band skin effect},
year = {2026},
howpublished = {\url{https://pith.science/paper/GJREFQ33}},
note = {Machine review of arXiv:2512.19745}
}
read the original abstract
Symmetry-protected ideal flat bands in one-dimensional (1D) Hermitian lattices are populated by compact localized states (CLS) - a special class of localization with wavefunctions confined within a small region. In this work, we discover that the non-Hermitian skin effect (NHSE) can appear in a flat band. Unlike conventional NHSEs for dispersive bands that are protected by nontrivial point-gap topology, the flat band remains a point on the complex-energy plane and is therefore always topologically trivial. We found that, intriguingly, the flat-band skin effect (FBSE) is associated with the non-trivial spectral topology of the dispersive bands enclosing the flat band on the complex-energy plane, so it only emerges within a finite range of non-Hermitian parameters and can counterintuitively disappear at large non-Hermiticity. Moreover, the gaps between the flat and the dispersive bands can close at higher-order exceptional points under both periodic and open boundary conditions. The flat-band wavefunctions are discontinuous in quantum distance across these exceptional points, signifying that the gap-closing is singular. The FBSE was experimentally observed in a non-Hermitian mechanical lattice. Our work reveals flat-band phenomena unique to non-Hermitian systems and highlights new possibilities in quantum geometry and localization control.
Forward citations
Cited by 2 Pith papers
-
Exceptional activated mode theory for generalized real-complex transitions
The real-to-complex spectral threshold is set by the lowest exceptional-point threshold among all pairs of reference modes, λ_c = Δ_act / G_act.
-
Generation of strongly localized skin solitons in non-Hermitian waveguide arrays with the Kerr effect
Analytical formulas derived via symbolic regression for soliton boundaries in non-Hermitian Kerr waveguide arrays, with continuum approximations showing NHSE-driven edge localization corroborated by numerics.
Reviewed August 3, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.