Pith. sign in

REVIEW 5 major objections 5 minor

Non-Hermitian Tight-Binding Bands in Graphene: Optical Conductivity, Strain Effects, and Bernal Bilayer Extension

T0 review · 5 major / 5 minor · reviewed 2026-08-04 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read This paper claims that a Möbius-transformed hopping integral, with a single anisotropy parameter B, extends graphene's tight-binding model into a non-Hermitian, geometry-dependent regime at B≠0 while reducing exactly to the standard model a

desk verdict At its calibrated optimum this model reduces to ordinary tight-binding, and the strained results rest on an unvalidated hybrid replacement, so the central benchmark claim is not supported. read the letter →

arxiv 2607.05470 v2 pith:V67J6ZCP submitted 2026-07-06 cond-mat.mtrl-sci

Maolin Bo , Yaorui Tan , Sunxin Fan , Xiang Chen , Yunhu Zhu , Zhongkai Huang , Chuang Yao This is my paper · ORCID
classification cond-mat.mtrl-sci PACS 73.22.Pr
keywords graphenetight-bindingmodelMöbiustransformationnon-Hermitianbandsnonreciprocalhoppinguniaxialstrainopticalconductivitybilayer
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper proposes that the hopping integral between nearest-neighbor π orbitals in graphene can be written as a Möbius transformation of the complex bond vector times an exponential distance decay, with one parameter, B, controlling the magnitude of bond-angle anisotropy. At B=0 the model is exactly the standard isotropic tight-binding result, and the paper's calibration against the conventional inverse-square distance scaling finds Bopt=0, so the model's nontrivial predictions all come from nonzero B. For B≠0 the three nearest-neighbor hoppings become distinct and complex, and because the reverse-bond hopping is no longer the complex conjugate of the forward one, the assembled Hamiltonian is non-Hermitian; the paper interprets the resulting imaginary energy bands as a nonreciprocity diagnostic rather than a lifetime. Under uniaxial strain, a hybrid distance-law × Möbius-angular scheme produces direction-dependent band gaps and optical dichroism, while the Dirac cone stays closed in Hermitian assembly and the bilayer bias-gap relation ΔK≈2δ is preserved. A sympathetic reader would care because this offers a minimal, reproducible single-particle benchmark for exploring non-Hermitian and anisotropic effects in graphene before adding electron-correlation physics.

What carries the argument

The load-bearing object is the Möbius-transformed hopping φ(w)=(Aw+B)/(Cw+D), specialized to C=1, D=0, so φ(w)=A+B/w, where w=X+iY is the complex coordinate of the bond vector. This single expression does three jobs: for B≠0 it makes the three nearest-neighbor hoppings bond-angle dependent and complex; it breaks the reverse-bond identity t(-dr)=t(dr)*, making the assembled Hamiltonian non-Hermitian; and through the ratio φ(w_j)/φ_eq(w_j) it provides the angular modulation multiplied onto a distance-decay law to describe strained graphene. The exponential factor exp[-(r-r0)/λ] supplies the length dependence, and the choice between Hermitian and non-Hermitian spectra is dictated by whether the

What would settle it

A density-functional or Wannier-interpolation calculation of the three nearest-neighbor hopping integrals of graphene at 5% uniaxial strain, checked against the model's predicted A+B/w angular dependence, would settle it: if the best-fit B is statistically zero, the anisotropic and nonreciprocal bands are artifacts of the parametrization.

Watch

Extended reading notes

Core claim

Graphene's nearest-neighbor hopping is parametrized as t(dr)=((Aw+B)/(Cw+D)) exp[-(r-r0)/λ], with C=1, D=0 and w the complex coordinate of the bond vector, so t=A+B/w. B is the dial: B=0 gives the standard isotropic -2.8 eV hopping and Hermiticity; B≠0 makes the hoppings distinct and complex, and φ(-w)=A-B/w ≠ φ(w)*=A+B/w, so t(-dr)≠t(dr)*. The asymmetric forward/backward structure factors make H(k) non-Hermitian, with complex eigenvalues read as nonreciprocity, not lifetime. Even at B=-3, Hermitian assembly keeps the K-point Dirac cone closed; the hybrid distance×angle scheme opens direction-dependent gaps, the bilayer bias relation ΔK≈2δ survives, and optical conductivity gives the univers

Load-bearing premise

The load-bearing premise is that the true π-hopping integral's dependence on bond angle and strain really factors into a directional part of the specific form A + B/w (with a single real B controlling both anisotropy and nonreciprocity) times an exponential or power-law distance decay; the paper gives no microscopic derivation of this factorization and its own calibration only supports B=0.

Editorial extensions

If this is right

  • At B=0 the model is a faithful benchmark: it matches standard tight binding and the inverse-square distance scaling exactly, so it can serve as a controlled starting point for non-Hermitian extensions.
  • Hermitian assembly of the geometry-dependent hoppings preserves the gapless Dirac cone at K for any B, so bond-angle anisotropy alone does not open a gap in unstrained graphene.
  • Nonzero B turns on a nonreciprocal hopping channel whose imaginary band structure (electronvolt-scale lobes along Γ–K–M) is a diagnostic of non-Hermitian assembly, not a lifetime effect.
  • The hybrid strain scheme with B≠0 predicts direction-dependent gaps that grow with |B| (e.g., ~1.35 eV geometric and ~3.16 eV hybrid at 24% armchair strain, B=-3), giving a tunable knob for strain engineering.
  • The same intralayer model embedded in a bilayer yields ΔK≈2δ for the interlayer bias (δ=50 meV gives ΔK≈0.10 eV), independent of B, and the optical conductivity reproduces the universal σ0 plateau with strain-induced dichroism.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Because Bopt=0 at equilibrium, the paper's nonzero-B predictions are scans of an unconstrained parameter; a decisive test would be to fit B to a measured or calculated strain-dependent gap, which would separate the model's predictive content from its illustrative range.
  • The non-Hermitian imaginary bands, though formally an assembly artifact of the A+B/w parametrization, may map onto real nonreciprocal transport in laser- or substrate-driven graphene; the predicted k-space lobe structure is a target for future coupled-mode or Floquet calculations.
  • The same Möbius factor could be transferred to other honeycomb or anisotropic lattices by fitting A, B, and λ to first-principles hopping integrals; the model's one-knob simplicity is its main advantage for such extensions.
  • The strain-optical dichroism prediction is most testable in the few-percent strain window accessible in experiments; the 12–24% scans are extrapolations and should not be compared directly to current measurements.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

5 major / 5 minor

Summary. This paper proposes a geometry-dependent nearest-neighbor hopping parameterization for graphene π bands, t(dr) = ((Aw+B)/(Cw+D)) exp[-(r-r0)/λ], with C=1, D=0, called the Tan–Bo model. It compares B=0 with standard Slater–Koster/Harrison scaling, uses B=-3 as an illustrative anisotropic/non-Hermitian preset, constructs Hermitian and non-Hermitian Hamiltonians via forward/backward bond sums, embeds the model in a Bernal bilayer, studies uniaxial strain with a hybrid hopping rule, and computes Kubo optical conductivity. The paper claims the model is a reproducible single-particle benchmark and parameterization reference for future non-Hermitian correlated calculations.

Significance. If the Tan–Bo factorization were independently validated, it would offer a compact two-parameter (A,B) way to interpolate between isotropic TB and strongly anisotropic/nonreciprocal hopping, with potential value as a benchmark for correlated calculations. The B=0 limit correctly recovers the standard Dirac cone and SK hopping, and the bilayer ΔK ≈ 2δ check is robust. However, the calibration in §2.3 is circular (Bopt=0 by construction), the strained-regime results abandon the full model in favor of an unvalidated hybrid in §2.4, and the §3.4 gap numbers are internally inconsistent. As it stands, the significance of the anisotropic/non-Hermitian results is illustrative rather than evidential.

major comments (5)
  1. [§2.3] The calibration is circular: L(B) is minimized against t_SK plus a K-point gap penalty, so Bopt=0 is effectively imposed by the target. No independent data (DFT, ARPES, STM, or strain optics) fix B≠0. The factorization t(dr)=φ(w)exp[-(r-r0)/λ] is introduced phenomenologically in §2.1 without derivation from overlap integrals or a fitting justification. Since B=-3 is admitted to be illustrative, the non-Hermitian complex bands, Im(E) lobes, and optical dichroism are scans of a free parameter, not a validated benchmark. Please fit B to an external observable or explicitly reframe the paper as a toy-model study.
  2. [§2.4] The strain section abandons the defining model: substituting t=φ(w)exp[-(r-r0)/λ] into strain calculations is said to yield a 'nonphysical' gap on the order of electronvolts, so all strained results use the hybrid tj=(φ(w_j)/φ_eq,j)t_Pereira(r_j). This hybrid is not a limit of the Tan–Bo expression; at B=0 the Möbius ratio is discarded entirely (Table 2). Consequently Figs. 5–8 and 10 describe a different, ad hoc model. The hybrid needs a controlled derivation or independent validation before strain-induced gaps and dichroism can support the benchmark claim.
  3. [§3.4] The reported strain gaps are internally inconsistent. At 24% strain the text states 'zigzag and armchair gaps reached 54 and 0.98 eV' and then 'at B=−3, the armchair and zigzag gaps reached 0.818 and 1.35 eV'; the Figure 7 discussion later gives zigzag 2.416 eV and armchair 3.164 eV at B=-3, 24%. These numbers cannot all be correct. Please recompute, define the gap (direct at K vs indirect over the path), and ensure text and figures agree.
  4. [§2.2, §3.2] The non-Hermitian assembly is formal: t(-dr)≠t(dr)* yields complex bands, but no physical mechanism or observable is tied to Im(E). The optical section deliberately excludes non-Hermitian eigenvalues, and the paper states Im(E) is not a lifetime. Without a concrete prediction—e.g., direction-dependent transmission, an exceptional-point signature, or a specific coupling to a bath—the claim that these complex bands are 'meaningful' is unsupported. This is a load-bearing issue for the central benchmark claim.
  5. [§2.1] The claimed LCAO-to-Tan–Bo derivation chain is not shown. The text states the LCAO estimate t ≈ V_ppπ S and then simply defines t(dr) as a Möbius factor times an exponential. If the model is intended as a phenomenological parameterization, that is acceptable, but contribution (1) in the Introduction claims a derivation chain; this claim is unsupported as written.
minor comments (5)
  1. [Abstract] Typo: 'Tthe' should be 'The'.
  2. [General] Several equations are corrupted in the rendered text (Hamiltonian, strain tensor, dimensionless bond lengths l_j, K-gap definitions). These must be typeset correctly before publication.
  3. [§2.6/§3.4] Please define the K-point gap used in strain scans: is it the direct gap at K, or the minimum gap along the path? The text alternates between 'K-path gap' and 'bandgap', and the reported numbers are difficult to reproduce without this definition.
  4. [Figs. 5–8] Figure captions should clearly distinguish which model is shown: full Tan–Bo vs hybrid, and Hermitian vs non-Hermitian assembly. Current captions are ambiguous.
  5. [Conclusions] The abstract and conclusions would benefit from stating explicitly that B=-3 is an illustrative dial, not an optimum, to avoid implying a fitted parameter.

Circularity Check

1 steps flagged · score 2.0 of 10

Bopt=0/SK equivalence is a definitional calibration identity; central non-Hermitian, bilayer, and optical results are not circular.

  1. self definitional [§2.3 (Parameter Calibration), echoed in Abstract and §3.1]
    "With the Tan–Bo model parameters fixed at A = −2.8, C = 1, and D = 0, the objective function is minimized within the narrow interval B ∈ [−0.05,0.05]. ... L(B) = MSE(TanBo, SK) + 5000 ΔK(B). ... In an equilibrium structure, Bopt = 0 ... At this point, the Tan–Bo isotropy (B = 0) matches precisely with the standard TB/SK model."

    At equilibrium r = r0 and B = 0, the Tan–Bo hopping reduces to t = A·exp[−(r−r0)/λ] = A = −2.8 eV, while the SK hopping is Vppπ·(r0/r)^2 = Vppπ = −2.8 eV. Thus the MSE term in L(B) is identically zero at B = 0 by the definitions of A and of the objective, before any minimization. The reported Bopt = 0 and the 'equivalence' to the SK model are therefore not independent discoveries; they are the calibration target built into L(B). The paper explicitly labels this a calibration, so it is a definitional identity rather than a load-bearing prediction, but the abstract and conclusions present it as a revealed result.

full rationale

The only step that is literally circular is the §2.3 calibration identity: because A is preset to −2.8 eV and B = 0 removes the Möbius angular factor, the equilibrium Tan–Bo hopping is identical to the standard SK value by construction, and minimizing an objective whose leading term is the MSE against SK must select B = 0. The paper itself calls this 'SK calibration', so it is a self-consistency check, not a prediction. The central non-Hermitian construction is an explicit ansatz: H_NH is assembled from f_fwd and f_bwd, and the B ≠ 0 complex bands follow from the stated definition of φ(w) = A + B/w; the paper repeatedly labels B = −3 as an 'illustrative anisotropy dial' rather than a fitted optimum, which removes the 'fitted input called prediction' pattern. The Tan–Bo model is cited to the authors' own refs [23,24], but the equations are restated in this paper, so the self-citation is provenance for an ansatz, not an independent proof that is being leaned on to close an argument. The strain section is a limitation rather than circularity: §2.4 admits that the full Tan–Bo exponential formula gives 'a nonphysical gap on the order of electronvolts under strain', so the hybrid hopping tj = (φ/φeq)·t_Pereira is a new model definition, not a consequence of the proposed benchmark; this undermines external validity but does not make the algebra circular. The Bernal bilayer ΔK ≈ 2δ check and the Kubo optical-conductivity checks are computed from the stated Hamiltonians and do not reduce to the fitted B; they are independent internal consistency tests. Overall, one definitional calibration identity exists, but the main derivation chain remains non-circular; score 2 reflects that minor self-referential step without treating the model proposal as circular.

Assumptions & free parameters 6 free parameters · 7 assumptions · 0 invented entities

The paper's central model depends on the ad hoc Möbius ansatz and on a post-hoc hybrid strain scheme; only the B=0 limit, which is standard TB/SK, is independently grounded. The non-Hermitian interpretation and B≠0 effects rest on the authors' own construction with no external benchmark.

free parameters (6)
  • B (Möbius anisotropy/non-Hermiticity knob) = 0 (optimum); illustrative −0.75, −1.5, −3
    Central parameter in φ(w)=A+B/w. Bopt=0 is obtained by minimizing ℒ(B)=MSE vs SK + 5000·ΔK(B) over [−0.05,0.05] (§2.3). B≠0 values are hand-chosen to demonstrate strong anisotropy; no independent fit.
  • A (isotropic hopping amplitude) = −2.8 eV
    Set to the standard TB/SK value in §2.1; the model's B=0 limit is therefore identical to standard TB by construction.
  • λ (exponential decay length) = 0.5 Å
    Chosen by hand in §2.1; appears in t(dr)=φ(w)exp[-(r−r0)/λ], not calibrated and largely bypassed by the hybrid strain scheme.
  • C and D (Möbius denominator/pole parameters) = C = 1, D = 0
    Fixed by 'physical constraints' and parameter-redundancy arguments (§2.1); pole at w=0, K-point closure preserved. They are model parameters chosen ad hoc for simplicity.
  • η (Kubo Lorentzian broadening) = 0.12 eV
    Numerical broadening for optical conductivity (§2.7); tests at 0.05–0.15 eV shift the van Hove peak by <80 meV, so it is not load-bearing.
  • DOS Gaussian broadening σ = 0.06–0.08 eV
    Used in DOS calculations; affects peak heights but not central band-structure claims.
assumptions (7)
  • domain assumption Nearest-neighbor, one-orbital (2pz) TB is a sufficient single-particle description of the π bands of graphene and Bernal bilayer graphene.
    Invoked throughout; standard but limits validity for high-energy, optical, and correlated regimes.
  • ad hoc to paper The hopping integral can be written as φ(w)·exp[-(r−r0)/λ] with a Möbius φ(w).
    Eq. (2), central ansatz; no derivation from overlap integrals and no independent B≠0 fit.
  • ad hoc to paper Under strain, hopping factorizes as tj(ε) = (φ(wj)/φeq,j)·t_Pereira(rj), rather than the full Tan–Bo formula, because the full formula gives unphysical eV gaps (§2.4).
    Post-hoc hybrid scheme; not derived and only the ratio φ/φeq is used.
  • ad hoc to paper Non-Hermitian Hamiltonian with f_fwd and f_bwd assembled separately is a meaningful model of 'nonreciprocity' even with no dissipation.
    §2.2; complex eigenvalues are explicitly not lifetimes, so their physical status is asserted rather than derived.
  • domain assumption Slater–Koster/Harrison t(r)=Vppπ(r0/r)^2 and Pereira exponential t0·exp[−β(l−1)] describe the distance dependence.
    Used as reference and in hybrid strain scheme (§2.3, §2.6); from prior literature.
  • domain assumption Kubo–Greenwood formula with independent-particle TB velocities (no excitonic/self-energy corrections) gives meaningful optical conductivity.
    §2.7; authors acknowledge the UV peak is blueshifted relative to the ~4.6 eV exciton.
  • domain assumption McCann four-band bilayer Hamiltonian with γ1 = −0.39 eV and interlayer bias δ is valid.
    §2.5; standard model.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Non-Hermitian Tight-Binding Bands in Graphene: Optical Conductivity, Strain Effects, and Bernal Bilayer Extension." pith.science (2026). https://pith.science/paper/V67J6ZCP

@misc{pith2026260705470,
  author       = {Pith},
  title        = {Pith review of: Non-Hermitian Tight-Binding Bands in Graphene: Optical Conductivity, Strain Effects, and Bernal Bilayer Extension},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/V67J6ZCP}},
  note         = {Machine review of arXiv:2607.05470}
}
read the original abstract

Within the tight binding framework of graphenes {\pi} electron nearest neighbors, the Tan Bo model parametrizes transition energies t(dr) based on bond lengths and angles via the Mobius transformation combined with exponential decay. Comparisons between isotropic , geometrically anisotropic , and Slater Koster scales reveal that B = 0 is equivalent to the SK scheme, with L(B) reaching its optimum at Bopt = 0. The Hermitian assembly maintains the Dirac cone at the K point.The Tan Bo geometry dependent transition model and non Hermitian TB assembly scheme developed in this study provide a reproducible single particle benchmark and parameterization reference for future non Hermitian chemical calculations incorporating electron correlation effects in graphene systems.

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Pith tools

Reviewed August 4, 2026 · model on record in the stance chip above.