Chiral Magnetic Conductivity in the Tight-Binding Model of Dirac Semimetals

M.A. Zubkov; Mustafa Bohra; Yuexiang Zhang

Chiral Magnetic Conductivity in the Tight-Binding Model of Dirac Semimetals

T0 review · 3 major / 6 minor · reviewed 2026-08-01 · deepseek-v4-flash

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

Pith's one-line read The paper derives, from the lattice tight-binding model in the lowest Landau level, a closed-form longitudinal magnetoconductivity σ_zz = |e|^3 |B| v_F / (4π² ℏ ε) · (1−|M|²)/(2|M|²), and argues this confirms that the current is carried by

desk verdict Solid analytic Keldysh calculation with a genuinely new vertex-correction factor, but the CME-confirmation claim outruns the approximations and the key factor is sensitive to unquantified off-shell corrections. read the letter →

arxiv 2607.19589 v2 pith:B4XJSL6I submitted 2026-07-21 cond-mat.mes-hall hep-ph

Mustafa Bohra , Yuexiang Zhang , M.A. Zubkov This is my paper · ORCID

classification cond-mat.mes-hallhep-ph

keywords chiralmagneticeffectDiracsemimetallowestLandaulevelmagnetoconductivitySSHmodelimpurityscatteringvertexcorrectionKeldyshformalism

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

This paper tries to show that the negative longitudinal magnetoresistance of Dirac semimetals can be derived microscopically from the lattice tight-binding model, without invoking an equilibrium chiral chemical potential by hand. In a strong magnetic field the dynamics collapses onto the lowest Landau level, and the model becomes a one-dimensional SSH-type chain along the field. Impurity scattering is treated in the non-equilibrium Green's function technique, and the field-induced correction to the self-energy is resummed through the impurity ladder. The resulting closed formula for σ_zz contains a vertex factor (1−|M|²)/(2|M|²) that was missing from the earlier continuum calculation, and the paper argues that the current is carried by the chiral (axial) charge imbalance. If correct, it turns the heuristic chiral-magnetic-effect picture into a calculable lattice statement and explains why the conductivity plateaus with field before decreasing at very large fields.

What carries the argument

The load-bearing object is the lowest-Landau-level-projected effective Hamiltonian, which is exactly an SSH-type one-dimensional two-band model along the magnetic field: H_LLL(θ) = v_F [γ0γ3 sinθ + γ0 (m + 1 − cosθ + |B|/2)]. The transport argument proceeds through the non-equilibrium Green's function technique: the impurity self-energy is computed from the LLL propagator, the field-induced self-energy correction is calculated, and the repeated impurity scattering (the ladder) is resummed. After the on-shell reduction the ladder kernel becomes an algebraic map whose odd-channel eigenvalue is q = (1−|M|²)/(1+|M|²); the final conductivity is proportional to q/(1−q) = (1−|M|²)/(2|M|²). Here |M|

What would settle it

Evaluate the linearized ladder equation numerically with the full matrix self-energy width (not its on-shell projection) for |M| ≈ 0.1; if the odd-channel eigenvalue of the kernel differs from q = (1−|M|²)/(1+|M|²) by order εa/(ℏv_F), the closed form fails exactly where it departs from the earlier result.

Watch

Extended reading notes

Core claim

The central claim is that in the quantum limit of a strong magnetic field, a lattice-regularized Dirac semimetal has longitudinal conductivity σ_zz = |e|^3 |B| v_F / (4π² ℏ ε) · (1−|M|²)/(2|M|²), with ε = u_0² n_imp |eB|/(4π v_F ℏ²) the impurity half-width and |M| ≈ Δ/μ the normalized mass (plus small lattice and field corrections). This is derived rather than assumed: the lowest-Landau-level projection gives an SSH-type one-dimensional Hamiltonian; the impurity self-energy is evaluated explicitly; and the field-induced self-energy correction is resummed through the impurity ladder, which satisfies the Ward identity. The factor (1−|M|²)/(2|M|²) is the complete vertex correction, and the rela

Load-bearing premise

The whole result hinges on the claim that the repeated-impurity-scattering ladder can be reduced to one number q, because the final conductivity contains 1/(1−q), so even a small error in q changes the answer a lot.

Editorial extensions

If this is right

The vertex correction multiplies the earlier continuum result by (1−|M|²)/(2|M|²); at fixed μ and moderate B the explicit |B| cancels the impurity width ε ∝ |B|, producing a plateau in σ_zz(B) rather than linear growth.
Once |B| is large enough that the field-dependent part of |M| becomes significant, σ_zz decreases with field, giving a microscopic account of the high-field downturn observed in low-temperature magnetotransport.
The relation J_3 = sign(B) v_F ρ_5 survives the entire ladder resummation, so the ratio of longitudinal current to axial density is fixed by the Fermi velocity and the current is genuinely carried by the field-induced chiral imbalance.
In the massless continuum limit |M| → 0 the chirality relaxation time diverges and the chiral magnetic current is non-dissipative; on a lattice the finite relaxation persists through the band-edge connection of the two chiral branches.
Strictly at T = 0 the reduced one-dimensional channels are subject to Anderson localization, so the Drude-type formulas apply to samples shorter than the backscattering mean free path or in the presence of finite dephasing.

Reading between the lines

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

Extending the paper's logic, the small-|M| regime is where the 1/|M|² enhancement is largest and also where the on-shell approximation is least controlled; a full finite-width solution of the ladder equation may regularize the apparent divergence and is the natural next step.
The same ladder factor should persist at finite frequency up to ω ∼ 2ε_on; above that the kernel becomes frequency-dependent, so a frequency-dependent plateau in σ_zz(Ω) is a testable signature of the mechanism.
Because the vertex correction renormalizes only the common response denominator and not the J_3–ρ_5 proportionality, a separate probe of the axial density — numerical or experimental — would test the mechanism more directly than conductivity alone.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Desk Editor's Note

read the letter

You should know two things about this paper. It contains one genuinely new analytic result — the ladder-resummed vertex correction that renormalizes the LLL magnetoconductivity by (1-|M|^2)/(2|M|^2) relative to the authors' earlier continuum calculation [48] — and the calculation looks honest and mostly self-consistent. But the abstract's claim that the results 'confirm' the chiral magnetic effect is too strong, and the paper's own caveats (Anderson localization, on-shell reduction) are more restrictive than the headline suggests.

What is actually new: the LLL reduction to an SSH-type Hamiltonian is explicit; the impurity self-energy integrals in Sec. V are done analytically; and the vertex resummation in Appendix F gives a closed-form correction that is absent from the earlier literature. The treatment of the equilibrium CME vanishing is sound, and the authors openly state that [48] made a mistake — a good sign. The algebra I checked holds together. There are no fitted parameters beyond the disorder strength.

The soft spots. First, the technical one: the kernel eigenvalue q that controls the final factor is computed after dropping off-shell and matrix-width terms at relative order eps a/(hbar v_F). The final conductivity is proportional to q/(1-q) ~ 1/(2|M|^2), so an uncomputed shift delta-q changes the answer by delta-q/(2|M|^2). The stated hierarchy only guarantees |M|^2 >> eps a/(hbar v_F), not that the coefficient of delta-q is small, and the deviation from Ref. [48] is largest exactly at small |M|. That needs either a numerical inversion of the full kernel or a bound on the coefficient. Second, the localization caveat in Sec. VII B is not a side remark: at strictly T=0 the reduced 1D channels are Anderson-localized, so the Drude formula applies only to samples shorter than the backscattering length or with finite dephasing. That limitation belongs in the abstract. Third, the phrase 'confirm CME' overstates: the result is a plateau in B (cut off by band-edge step functions and by M(B) at high fields), not the linear-in-B growth usually associated with negative magnetoresistance, and there is no reconciliation with the linear-in-B Kubo results of [16-18]. Minor but real: reference [47] has a placeholder DOI and its content is not verifiable; that must be fixed.

Overall judgment: this is a serious, mostly rigorous calculation that deserves a referee's time. I would send it to peer review with a request to address the kernel-sensitivity point, soften the CME-confirmation language, and move the localization restriction to the abstract. My verdict is conditional, in line with the reader's.

Referee Report

3 major / 6 minor

Summary. The paper analyzes a 3D lattice-regularized Dirac semimetal in parallel electric and magnetic fields, in the strong-field/LLL limit. Projecting onto the lowest Landau level and expanding transverse lattice operators, the model is reduced to an effective 1D SSH-type Hamiltonian along the field axis (Sec. II, Eq. (24)). Using Keldysh Green functions and a Born treatment of point-like impurities, the authors compute the impurity self-energy (Sec. V), the electric-field-induced correction to the Green function including an exact resummation of the impurity ladder (Sec. VI and Appendices E, F), and then evaluate the axial charge density and longitudinal current (Sec. VII). The central results are Eq. (119)/(124): in the low-energy, leading on-shell approximation, σ_zz = |e|^3 |B| v_F/(4π^2 ℏ ε) × (1-|M|^2)/(2|M|^2), with ε = u0^2 n_imp |eB|/(4π v_F ℏ^2) and |M| ≈ Δ/μ, together with the proportionality J_3 = sign(B) v_F ρ_5 (Eq. (113)). The paper presents this as a microscopic confirmation that the magnetoconductivity is due to the chiral magnetic effect, and as a correction to Ref. [48], whose bare-bubble result is multiplied by the factor (1-|M|^2)/(2|M|^2).

Significance. If correct, the result is significant: it gives a lattice-microscopic, Keldysh-based derivation of the quantum-limit longitudinal magnetoconductivity of a Dirac semimetal with impurity scattering, including an explicit vertex correction, and it clarifies the status of the chiral magnetic effect as a nonequilibrium dissipative transport phenomenon. The derivation is transparent in several places: the LLL reduction, the impurity self-energy integral, the spectral identities of Appendix C, and the algebraic diagonalization of the leading on-shell kernel in Appendix F are all carried out in closed form. The paper also explicitly states its validity hierarchies and acknowledges important limitations, including the zero-temperature Anderson-localization caveat in Sec. VII B and the on-shell nature of the vertex reduction in Appendix F. These strengths make the paper a useful contribution regardless of whether the final factor (1-|M|^2)/(2|M|^2) survives more exact treatment. The main unresolved issue is quantitative: the ladder eigenvalue q, which controls the most novel part of the result, is computed only at leading on-shell order, and the correction to it enters a 1/(1-q) denominator

major comments (3)

[Appendix F / Eq. (91), Eq. (F1), Eq. (119)] The central new factor (1-|M|^2)/(2|M|^2) comes from resumming the ladder denominator 1-q, where q=(1-|M|^2)/(1+|M|^2). Appendix F reduces the integral kernel to the algebraic map (F1) by dropping off-shell, matrix-width, and regular corrections of relative order ε a/(ℏ v_F). The eigenvalue q is therefore known only up to an uncomputed shift δq = C ε a/(ℏ v_F). Since 1-q = 2|M|^2/(1+|M|^2), the relative error in σ_zz is δq/(1-q) ≈ C ε a/(ℏ v_F) (1+|M|^2)/(2|M|^2). The stated hierarchy 2|M|^2/(1+|M|^2) ≫ ε a/(ℏ v_F) only guarantees |M|^2 ≫ ε a/(ℏ v_F); without a bound on C, the correction is uncontrolled precisely in the small-|M| regime where the factor differs most from Ref. [48]. A numerical inversion of the full finite-width kernel—which is a one-dimensional integral kernel in p_3—would settle whether Eq. (119) survives; this is a concrete and feasible check.
[Sec. VII B (penultimate paragraph)] The manuscript explicitly states that at strictly T=0 the reduced 1D channels are Anderson-localized and that the Drude-type expressions apply only to samples shorter than the backscattering mean free path or in the presence of finite dephasing. This is an important limitation of the headline result: Eq. (119)/(124) is not the DC conductivity of an infinite 3D sample at T=0 in the thermodynamic limit, but a finite-sample or dephasing-limited formula. The paper is honest about this, but the abstract and conclusion present the result without this qualification. The authors should either state the finite-sample/dephasing domain explicitly in the abstract and conclusion, or provide a localization-length estimate that shows the regime is experimentally relevant.
[Sec. VI / Eq. (92)] The solution of the linearized Schwinger-Dyson equation uses the on-shell projected drive and the algebraic kernel spectrum (91). The retarded/advanced components of the field-induced self-energy are computed in Eq. (E6) and are O(ε E_3/μ), which is consistent. However, the ladder resummation itself is justified only by the leading on-shell approximation; the regular parts of the kernel, which are not computed, are asserted to be of relative order ε a/(ℏ v_F). Since the pole at q=1 in the odd channel is the mechanism that produces the large factor 1/(2|M|^2), this is a load-bearing point. The manuscript would be materially strengthened by either a direct estimate of the coefficient C in the shift of q, or by numerically solving the full kernel for representative parameters.

minor comments (6)

[General] There are several typographical and grammatical errors: 'dirves' in the Introduction, 'Theoy' in Ref. [69], 'matirx' in Appendix A, 'Prove of it' in Appendix A. These should be corrected.
[Eq. (115) and Eq. (119)] The final conductivity formulas display only the lower band-edge step θ(|μ| - v_F(m+|B|a/2)), whereas the earlier expression (114) contains both lower and upper band-edge factors. For μa/v_F ≪ 1 the upper step is redundant, but this should be stated to avoid apparent inconsistency.
[Eq. (81)] The definition of D is written as 'D = (2 + 2ma + |B|a^2)/2' only in passing; later D is used again with a=1. Please define D once, consistently in both conventions.
[Sec. III / Eq. (35)] The paper says 'small temperature limit' in the abstract, but the calculation uses T=0 via n'(ω) = -δ(ω-μ). The finite-temperature extension is only mentioned as future work. Please make the T=0 assumption explicit in the abstract.
[Eq. (113)] The proportionality J_3 = sign(B) v_F ρ_5 is quoted as a direct relation, but the equation itself carries O(θ_F^2, ma, |B|a^2, ε/μ) corrections. In the conclusion this is stated without caveats; adding the same qualifiers would be more accurate.
[Appendix C] Appendix C explicitly states that the matrix chain rule for ∂_{p_3} δ(p_0-H) is not exact on the lattice because [H,H']≠0, and that the result is a leading on-shell identity with O(ε/μ) remainders. This is an important honesty check and should be highlighted more prominently in the main text, not only in the appendix.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the conductivity is computed from the lattice model; the CME form is a post-hoc rewriting, not an input.

full rationale

The headline result, Eqs. (119) and (124), is derived from the lattice Dirac Hamiltonian by LLL projection, an explicit Born impurity self-energy, and a resummed impurity ladder. The heuristic CME formula, Eq. (5), appears only as introductory motivation and is not used to compute the response; the response time tau_5,resp is explicitly defined after the calculation (Sec. VII B: "the algebraic identity rho_5 = (E.B) tau_5,resp/(2 pi^2) defines the response time tau_5,resp"), so the apparent CME proportionality is a rewriting of the already-derived sigma_zz and rho_5, not an assumed input. The vertex factor (1-|M|^2)/(2|M|^2) follows algebraically from the kernel eigenvalue q=(1-|M|^2)/(1+|M|^2), Eq. (91), and the geometric series in Eq. (92); no parameter is fitted to the predicted conductivity. Self-citations such as [43] and [48] provide context or a corrected comparison result, but the central derivation is self-contained rather than resting on an unverified self-cited premise. The paper itself flags the weak-disorder on-shell approximation and the Anderson-localization limitation in Sec. VII B ("strictly at T=0 the reduced one-dimensional channels are subject to Anderson localization..."); these are validity restrictions, not circular reductions. Overall, the derivation chain does not reduce by construction to its inputs.

Assumptions & free parameters 1 free parameters · 7 assumptions · 0 invented entities

The central claim rests on a hierarchy of stated limits rather than on fitted numbers: |B|a^2 ≪ 1 for the transverse expansion (Sec. II), mu a/v_F ≫ ma + |B|a^2 and the weak-disorder inequality for the on-shell ladder (Sec. VI), and the sample-scale/decoherence condition for Drude validity (Sec. VII B). The one model parameter with arbitrary value on which the final number depends is the disorder correlator u0^2 n_imp; the paper performs no fit to data, so no quantitative prediction of sigma's magnitude is made without specifying it. No new entities (particles, forces, conserved charges) are introduced; the heuristic mu_5 of the introduction is explicitly de-emphasized as an ill-defined construct.

free parameters (1)

disorder strength u0^2 n_imp = not fitted (model input)
White-noise correlator of Eq. (58); the final sigma_zz is inversely proportional to u0^2 n_imp (Eqs. 124-125). The paper makes no quantitative comparison, so the magnitude of the prediction is undetermined until this parameter is specified.

assumptions (7)

domain assumption Transverse continuum expansion: sin(Π_{x,y}) ≈ Π_{x,y}, 1−cos(Π_{x,y}) ≈ Π^2_{x,y}/2, valid for |B| ≪ 1/a^2 (magnetic length much larger than lattice spacing), while the longitudinal dispersion is kept fully lattice-valued
Sec. II Eqs. (14)-(18); the controlled limit that makes the LLL reduction exact-algebraic. Requires the stated hierarchy a ≪ l_B ≤ sqrt(2) l_mu.
domain assumption White-noise delta-correlated scalar disorder with ⟨U(x)U(y)⟩ = u0^2 n_imp δ^{(3)}(x−y)
Eq. (58); the entire dissipation mechanism is this single correlator; no other scattering channel is included.
domain assumption At T=0 only elastic impurity scattering remains; phonons vanish and Coulomb interactions are neglected
Sec. I and Conclusion: 'Coulomb interactions may still remain relevant, but we do not consider their effect.'
ad hoc to paper Born/weak-disorder + leading on-shell approximation: the impurity-ladder kernel in App. F reduces to the algebraic map (F1) with spectrum (91), dropping off-shell, matrix-width, and regular corrections of relative order eps a/(hbar v_F)
Appendices E and F; Sec. VI validity condition 2|M|^2/(1+|M|^2) ≫ eps a/(hbar v_F). This is the load-bearing technical premise for the vertex-corrected factor.
ad hoc to paper Spectral-decomposition and integration-by-parts identities in App. C hold at leading on-shell order; interband remainders are O(eps/mu)
Appendix C, paragraph after Eq. (C8): 'It is not an exact matrix-chain-rule result. As ρ5 = iE3|B|I/2, this determines the leading contribution.'
domain assumption DC linear response: the electric field is treated to first order via the Wigner gradient expansion; Zener tunneling between the split bands is exponentially suppressed
Conclusion: 'the above expressions disregard tunneling phenomenon... exponentially suppresses the effect due to nonzero gap.'
domain assumption The reduced 1D channels conduct Ohmically at DC, i.e., samples are shorter than the backscattering mean free path or have finite dephasing, despite strict T=0 Anderson localization
Sec. VII B, penultimate paragraph: 'strictly at T=0 the reduced one-dimensional channels are subject to Anderson localization... The Drude-type expressions above refer to the samples shorter than this scale, or to the presence of a finite dephasing.'

how reviews work

0 comments

Cite this review

Pith. "Pith review of Chiral Magnetic Conductivity in the Tight-Binding Model of Dirac Semimetals." pith.science (2026). https://pith.science/paper/B4XJSL6I

@misc{pith2026260719589,
  author       = {Pith},
  title        = {Pith review of: Chiral Magnetic Conductivity in the Tight-Binding Model of Dirac Semimetals},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/B4XJSL6I}},
  note         = {Machine review of arXiv:2607.19589}
}

read the original abstract

We consider the typical tight - binding model of Dirac semimetal in the presence of both magnetic and electric fields. The electric conductivity reveals dependence on magnetic field. We calculate this dependence in the limit of strong magnetic field, when the given model is described effectively by the one - dimensional SSH model because the dynamics in the plane orthogonal to magnetic field is reduced to that of the lowest Landau level (LLL). Considering the small temperature limit we take into account dissipation due to scattering on impurities. The corresponding dissipation rate is calculated explicitly. The obtained results confirm that the source of the magnetoconductivity in this system is chiral magnetic effect.

Discussion (0). Continue with ORCID to comment.

Pith tools

Reviewed August 1, 2026 · model on record in the stance chip above.