Pith. sign in

REVIEW 2 major objections 4 minor 27 references

Improved local spectral gap thresholds for lattices of finite dimension

T0 review · 2 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read Local spectral gaps of gapless spin lattices must shrink quadratically in the shortest side length.

desk verdict Improves the local gap threshold to optimal O(1/t^2) in any finite dimension, but the main proof leans on a sketched claim that needs a real proof. read the letter →

arxiv 1909.01516 v1 pith:ZALAH2QQ submitted 2019-09-04 quant-ph cond-mat.str-el

Anurag Anshu This is my paper · ORCID
classification quant-phcond-mat.str-el
keywords spectralgapfrustration-freeHamiltonianlocalthresholdcoarse-graineddetectabilitylemmaquantumspinlatticegaplesssystemsChebyshevpolynomials
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 establishes a tight quantitative link between the global spectral gap of a frustration-free local Hamiltonian (one whose ground state is annihilated by each local interaction) on a fixed-dimensional lattice and the spectral gaps of the same Hamiltonian restricted to finite hyper-rectangular blocks. The main theorem says the minimum local gap $\gamma(t_1,\dots,t_D)$ is bounded above by a constant, depending on dimension and interaction structure, times the global gap $\gamma$, plus a term of order $1/\min_q t_q^2$. For a gapless system, $\gamma \to 0$, so every $t_1 \times \dots \times t_D$ block must have a gap at most $O(1/\min_q t_q^2)$; the paper argues this quadratic decay is unavoidable, up to dimension-dependent constants, via parallel chains of a simple ferromagnetic model. This sharpens earlier local-gap thresholds in several dimensions and matters because local gaps are a practical route to proving global spectral gaps, which in turn govern correlation decay and phase structure.

What carries the argument

The load-bearing object is the coarse-grained Hamiltonian. On a chain, the lattice is divided into overlapping blocks $S_k$ and $T_k$ of length $t$, and $\hat{H}(t) = \sum_k (Q_{S_k} + Q_{T_k})$ is formed, where each $Q$ is the projector onto excited states of the block-restricted Hamiltonian; $\hat{H}(t)$ has the same ground space as the original Hamiltonian. The proof couples three tools: the detectability lemma, which controls powers of the layered product of local projectors on excited states; a low-degree Chebyshev 'step' polynomial, which shrinks the excited spectrum enough to yield $\gamma(\hat{H}(t)) \ge \frac{t^2\gamma}{400L^2g^2 + 3t^2\gamma}$; and the inequality $\gamma(\hat{H}(t)) \le 2\gamma/\gamma(t)$. Combining these gives the one-dimensional bound $\gamma(t) \le 103 L^2 g^2/t^2 + 6\gamma$. Higher dimensions are handled by viewing a slab Hamiltonian as a chain of 'column' Hamiltonians and applying the one-dimensional argument recursively, once per dimension.

What would settle it

A direct falsifier would be a family of translationally invariant frustration-free local Hamiltonians on a $D$-dimensional lattice, gapless as the system grows, for which the minimum spectral gap over $t \times \dots \times t$ regions decays faster than $1/t^2$, for example $\Theta(1/t^{5/2})$; this would break the claimed optimal additive term in Theorem C.1. A more surgical check is to test the polynomial-insertion identity (Equation 15) numerically for small $t$ just above $8L^2$: any failure there would invalidate the chain of inequalities that produces the quadratic threshold.

Watch

Extended reading notes

Core claim

The central discovery is Theorem C.1: for a frustration-free Hamiltonian $H = \sum_\alpha P_\alpha$ made of local projectors on a $D$-dimensional lattice whose interaction graph has degree $g$ and whose terms can be partitioned into $L$ commuting layers, the minimum spectral gap over hyper-rectangles of side lengths $t_1,\dots,t_D$ satisfies $\gamma(t_1,\dots,t_D) \le 6D\gamma + 200 L^2 g^2 6^D / \min_q t_q^2$, provided each side length is larger than a constant depending on $L$ and $g$. In the gapless limit this gives $\gamma(t,\dots,t) = O(1/t^2)$, and the example of many independent Heisenberg ferromagnetic chains shows the $1/t^2$ term is necessary even for translationally invariant systems, so the dependence on $\min_q t_q^2$ is optimal up to the dimension-dependent constant. The theorem applies to open and periodic boundary conditions and does not require translation invariance.

Load-bearing premise

The proof's load-bearing step is the assertion that a low-degree polynomial of the full detectability operator can be inserted between the block projectors of the coarse-grained Hamiltonian without changing the operator; this absorption step is justified in the paper only by an outline and depends on the overlap between the two block families being at least a quarter of the block length.

Editorial extensions

If this is right

  • Any gapless frustration-free Hamiltonian on a fixed-dimensional lattice must have local spectral gaps over side-length $t$ regions that are at most $O(1/t^2)$; no such model can keep those region gaps at $\Theta(1/t)$.
  • For finite systems, the theorem gives a finite-size criterion: if every $t_1 \times \dots \times t_D$ region has a gap larger than the stated threshold, the global Hamiltonian is guaranteed gapped.
  • The bound holds without translation invariance and for open and periodic boundary conditions, so it applies to boundary-modified and disordered frustration-free systems.
  • Since the $1/\min_q t_q^2$ rate is optimal, further improvement in this type of threshold can only come from the dimension-dependent constant or from additional symmetry of the Hamiltonian.

Reading between the lines

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

  • A fully rigorous proof of the absorption identity for all admissible $t$ would make the quadratic threshold unconditional for arbitrary local interactions; numerical checks on small chains could reveal whether the identity is exact or only approximate outside the sketched overlap regime.
  • The same machinery might yield quantitative decay-of-correlation or entanglement bounds in gapless frustration-free systems, since it already converts global gap information into local spectral information with sharp scaling.
  • If the paper's conjecture for isotropic translationally invariant Hamiltonians is correct, the relevant quantity becomes the inverse-squared diameter $1/\sum_q t_q^2$, so long thin regions would be far less constrained than the shortest-side bound suggests; columnar quasi-one-dimensional models could test this.
  • The exponential dependence of the constant on dimension is an artifact of the recursive column-row decomposition, and a direct $D$-dimensional coarse-graining argument could plausibly reduce it to polynomial in $D$.
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

2 major / 4 minor

Summary. The paper studies frustration-free local Hamiltonians on finite-dimensional lattices and proves a local spectral gap threshold. The main result (Theorem C.1) states that for a frustration-free Hamiltonian on a D-dimensional lattice with spectral gap γ, the minimum spectral gap γ(t_1,...,t_D) over all hyper-rectangular regions of side lengths t_q satisfies γ(t_1,...,t_D) ≤ 6^D γ + 200 L^2 g^2 6^D / min_q t_q^2, with constants depending on the lattice dimension and the interaction's locality and layer structure. The proof uses a one-dimensional reduction: a coarse-grained Hamiltonian of overlapping intervals (S and T regions), an estimate on low-degree Chebyshev polynomials, and the detectability lemma. A recursive dimensional-reduction argument then extends the one-dimensional bound to D dimensions. The paper also presents a lower-bound example (parallel copies of the Heisenberg ferromagnet chain) showing that the quadratic decay in the shortest side length is optimal up to dimension-dependent constants, and an example showing that a version with an average over regions would not hold.

Significance. If fully established, this result is a significant step in the finite-size criteria literature for frustration-free Hamiltonians: it improves the prior local gap thresholds for D-dimensional lattices (from O(log^2 t / t) in [KL18] and O(1/t) in [Lem19a]) to the optimal O(1/t^2) scaling, with a simple proof technique based on coarse-grained Hamiltonians and the detectability lemma. The lower-bound example correctly demonstrates that the scaling in the shortest side length cannot be improved. However, the manuscript as written is not complete: the key identity behind the main theorem is only sketched, and there is an arithmetic error in the derivation of the constant in Theorem B.1. These issues must be fixed before the claims can be accepted.

major comments (2)
  1. [B.3] The derivation of the constant in Theorem B.1 is incorrect. The displayed bound γ(¯H(t)) ≥ t^2γ/(400 L^2 g^2 + 3 t^2γ) combined with Eq. (8), i.e., 2γ/γ(t) ≥ γ(¯H(t)), gives γ(t) ≤ 800 L^2 g^2 / t^2 + 6γ, not the stated 103 L^2 g^2 / t^2 + 6γ. The constant 103 appears in the theorem statement and is then propagated through the recursion in Theorem C.1. Since the paper explicitly says the constants are not optimized, this error is repairable, but as written the proof does not establish the stated numerical bound.
  2. [E, Claim E.1] Claim E.1 is load-bearing for Lemma B.5 and hence for Theorems B.1 and C.1, but its proof is only an outline. The claim that all layer operators DL_α can be 'absorbed' into the S and T projectors because the overlap between adjacent S and T sets is at least ⌊t/4⌋ is not a valid inference: the DL_α are products over entire layers and do not commute with the projectors (1-Q_Sk) and (1-Q_Tk), so spatial overlap alone does not imply the identity in Eq. (15). For example, in the L=2 case with F(x)=x, the identity would require (1-Q_S)(1-P_2)(1-P_1)(1-P_2)(1-Q_T) = (1-Q_S)(1-Q_T), which is not a formal consequence of the assumptions. The author should provide a complete proof of Claim E.1 or replace the upper bound in Lemma B.5 with a fully rigorous argument; a citation to [AAG19, Claim B.1] is not sufficient in a manuscript that promises a proof for completeness.
minor comments (4)
  1. [C.1] The condition in Theorem C.1 is garbled: 'Suppose 264DL<t s<n s/ 5' should presumably be a condition of the form t_s > c_D L and t_s < n_s/5 for an explicit dimension-dependent constant c_D. Please correct the typesetting.
  2. [C.1] In the proof of Theorem C.1, the base-case display '10 3 2Lg' appears to be a misprint for '103 L^2 g^2', and similar notation errors make the constant bookkeeping hard to follow.
  3. [B.5] The lower bound in Lemma B.5 is stated to follow from Lemma B.4, but the connection is not made explicit: the coarse-grained Hamiltonian has two layers (the S projectors and the T projectors), so a short explanation of why Lemma B.4 applies would improve readability.
  4. [2.1] In the two-dimensional proof outline, the sentence 'The overall additive factor is O(1/t_1^2 + 1/t_1^2)' should read O(1/t_1^2 + 1/t_2^2); the repeated t_1 is a typo.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the local-gap threshold derives from prior lemmas and the self-cited technical claims are independent of the target result.

full rationale

The derivation chain is not circular. Theorem B.1 is proved by combining the variational inequality gamma(bar H(t)) <= 2 gamma / gamma(t) (Eq. 8), the detectability-lemma bounds of Lemmas B.2 and B.3, the Chebyshev estimate Claim A.1, and the layer-absorption identity in Claim E.1. None of these inputs is a restatement of the conclusion gamma(t) <= 6 gamma + O(1/t^2); in particular, Eq. 8 is derived from the spectral inequalities defining the coarse-grained Hamiltonian, not from the target bound. The self-citations to [AAV16] and [AAG19] are to prior technical lemmas whose stated assumptions do not contain the local-gap threshold, so they are independent support under the review rules. Claim E.1 is admittedly given only as a proof outline and is cross-referenced to [AAG19, Claim B.1]; this is a rigor/completeness concern of the kind raised by the skeptic's noncommutativity objection, not a circular reduction, because identity (15) is not equivalent by construction to the threshold being proved and no fitted parameter is later renamed as a prediction. The optimality example using parallel Heisenberg chains explicitly computes gamma(t,n2,...,nD) = pi^2/(2t^2) and is an independent witness rather than an assumption of the theorem. Overall, no load-bearing step reduces to its own input.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

No free parameters are fitted; the theorem is a pure derivation. The proof rests on standard quantum information lemmas (detectability lemma, Jordan's lemma) and on structural assumptions about the Hamiltonian's interaction graph. No new physical entities are introduced.

assumptions (5)
  • standard math Detectability lemma and its converse (Lemmas B.2, B.3, B.4)
    Used in Lemma B.5 to bound the coarse-grained detectability operator DL(t) via a polynomial in DL(H)†DL(H); taken from Aharonov et al., Gao, and AAV16.
  • domain assumption The coarse-grained Hamiltonian Hbar(t) has the same ground space G as H
    Assumed in §B.2 for the coarse-grained construction; follows from frustration-freeness and the definition of projectors Q_S, but is not re-proven in full generality.
  • domain assumption The local projectors P_ij can be divided into L layers where terms within a layer commute, and each term non-commutes with at most g others
    Assumed in §B; for hypercubic lattices the authors state L,g ≤ (3D)^D in Section C.
  • standard math Jordan's lemma for pairs of projectors
    Used in Appendix D to prove the converse detectability bound for L=2.
  • standard math Chebyshev polynomial estimate (Claim A.1)
    Derived in Appendix A with explicit bounds that the proof of Theorem B.1 relies on.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Improved local spectral gap thresholds for lattices of finite dimension." pith.science (2026). https://pith.science/paper/ZALAH2QQ

@misc{pith2026190901516,
  author       = {Pith},
  title        = {Pith review of: Improved local spectral gap thresholds for lattices of finite dimension},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZALAH2QQ}},
  note         = {Machine review of arXiv:1909.01516}
}
read the original abstract

Knabe's theorem lower bounds the spectral gap of a one dimensional frustration-free local hamiltonian in terms of the local spectral gaps of finite regions. It also provides a local spectral gap threshold for hamiltonians that are gapless in the thermodynamic limit, showing that the local spectral gap much scale inverse linearly with the length of the region for such systems. Recent works have further improved upon this threshold, tightening it in the one dimensional case and extending it to higher dimensions. Here, we show a local spectral gap threshold for frustration-free hamiltonians on a finite dimensional lattice, that is optimal up to a constant factor that depends on the dimension of the lattice. Our proof is based on the detectability lemma framework and uses the notion of coarse-grained hamiltonian (introduced in [Phys. Rev. B 93, 205142]) as a link connecting the (global) spectral gap and the local spectral gap.

Figures

Figures reproduced from arXiv: 1909.01516 by the authors.

Figure 1
Figure 1. (a) We can view the hamiltonian on two dimensional lattice as a hamiltonian on one dimensional chain of column of spins (dark blue rectangles). The interaction H4 between columns 4 and 5 is shown as the red rectangle, which decomposes as H4 = Pn2−1 j=1 P4,j . (b) Our strategy is to lower bound the spectral gap of H with the spectral gap of hamiltonian hS supported on the red region S × {1, 2, . . . n2}. The spectral… view at source ↗
Figure 2
Figure 2. Dividing the chain into contiguous segments of length t: Here, we assume n = 37 and t = 5. The remainder when n is divided by t is 2. We set r1 = r2 = 1 and rk = 0 for k > 2. The green rectangles represent the sets Sj . The red and the blue rectangles represent the sets Tj . The blue rectangles are to be viewed as a single contiguous region on the closed chain when Hn 6= 0 and are assumed to not exist on the open ch… view at source ↗
Figure 3
Figure 3. Assume n = 38, t = 18 and Hn = 0 (open chain). In this case, r = 2. There is exactly one set T1 and two sets S1, S2. Lemma B.5. It holds that 1 − 3γ(H¯ (t)) ≤ maxψ∈G⊥ kDL(t)|ψi k2 ≤ maxx∈(0,1− γ g2+γ )Step t 8L , γ g2+γ (x). Now, we proceed to the proof of our main theorem. B.3 Proof of Theorem B.1 We start with the inequality for all 1 ≤ k ≤ quo QSk + QTk  1 γ(Sk) hSk + 1 γ(Tk) hTk  1 γ(t) (hSk + hTk ). Note that… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

27 extracted references · 22 canonical work pages

  1. [1]

    Entanglement subvolume law for 2D frustration-free spin systems

    Anurag Anshu, Itai Arad, and David Gosset. Entanglement subvolume law for 2d frustration-free spin systems, 2019. https://arxiv.org/abs/1905.11337

    work page Pith review arXiv 2019

  2. [2]

    The detectability lemma and quantum gap amplification

    Dorit Aharonov, Itai Arad, Zeph Landau, and Umesh Vazirani. The detectability lemma and quantum gap amplification. In Proceedings of the Forty-first Annual ACM Symposium on Theory of Computing , STOC '09, pages 417--426, New York, NY, USA, 2009. ACM

    work page 2009

  3. [3]

    Simple proof of the detectability lemma and spectral gap amplification

    Anurag Anshu, Itai Arad, and Thomas Vidick. Simple proof of the detectability lemma and spectral gap amplification. Phys. Rev. B , 93:205142, May 2016

    work page 2016

  4. [4]

    An area law and sub-exponential algorithm for 1d systems

    Itai Arad, Alexei Kitaev, Zeph Landau, and Umesh Vazirani. An area law and sub-exponential algorithm for 1d systems. http://arxiv.org/abs/1301.1162, 2013

    arXiv 2013

  5. [5]

    An improved 1d area law for frustration-free systems

    Itai Arad, Zeph Landau, and Umesh Vazirani. An improved 1d area law for frustration-free systems. Phys. Rev. B. , 85, 2012

    work page 2012

  6. [6]

    A class of two-dimensional aklt models with a gap, 2019

    Houssam Abdul-Rahman, Marius Lemm, Angelo Lucia, Bruno Nachtergaele, and Amanda Young. A class of two-dimensional aklt models with a gap, 2019. arXiv preprint arXiv:1901.09297

    arXiv 2019

  7. [7]

    Undecidability of the spectral gap in one dimension, 2018

    Johannes Bausch, Toby Cubitt, Angelo Lucia, and David Perez-Garcia. Undecidability of the spectral gap in one dimension, 2018. arXiv preprint arXiv:1810.01858

    arXiv 2018

  8. [8]

    Gapped and gapless phases of frustration-free spin-12 chains

    Sergey Bravyi and David Gosset. Gapped and gapless phases of frustration-free spin-12 chains. Journal of Mathematical Physics , 56(6):061902, 2015

    work page 2015

Show all 27 references
  1. [9]

    Fernando G. S. L. Brand \ a o, Aram W. Harrow, and Micha Horodecki. Local random quantum circuits are approximate polynomial-designs. Communications in Mathematical Physics , 346(2):397--434, Sep 2016

    2016

  2. [10]

    Cubitt, David Perez-Garcia, and Michael M

    Toby S. Cubitt, David Perez-Garcia, and Michael M. Wolf. Undecidability of the spectral gap. Nature , 528:207 -- 211, Dec 2015. Article

    2015

  3. [11]

    Eldar and A

    L. Eldar and A. W. Harrow . Local hamiltonians whose ground states are hard to approximate. In 2017 IEEE 58th Annual Symposium on Foundations of Computer Science (FOCS) , pages 427--438, Oct 2017

    2017

  4. [12]

    J. Gao. Quantum union bounds for sequential projective measurements. Physical Review A , 92(5):052331, 2015

    2015

  5. [13]

    Correlation length versus gap in frustration-free systems

    David Gosset and Yichen Huang. Correlation length versus gap in frustration-free systems. Phys. Rev. Lett. , 116:097202, Mar 2016

    2016

  6. [14]

    Local gap threshold for frustration-free spin systems

    David Gosset and Evgeny Mozgunov. Local gap threshold for frustration-free spin systems. Journal of Mathematical Physics , 57(9):091901, 2016

    2016

  7. [15]

    Personal communication

    David Gosset, 2019. Personal communication

    2019

  8. [16]

    M. B. Hastings. Lieb-schultz-mattis in higher dimensions. Phys. Rev. B. , 69, 2004

    2004

  9. [17]

    M. B. Hastings. An area law for one dimensional quantum systems. J. Stat. Mech. , P08024, 2007

    2007

  10. [18]

    Hastings and Tohru Koma

    Matthew B. Hastings and Tohru Koma. Spectral gap and exponential decay of correlations. Communications in Mathematical Physics , 265(3):781--804, Aug 2006

    2006

  11. [19]

    Essai sur la g\` e om\` e trie \' a n dimensions

    Camille Jordan. Essai sur la g\` e om\` e trie \' a n dimensions. Bulletin de la S. M. F. , 3:103--174, 1875

  12. [20]

    Divide and conquer method for proving gaps of frustration free hamiltonians

    Michael J Kastoryano and Angelo Lucia. Divide and conquer method for proving gaps of frustration free hamiltonians. Journal of Statistical Mechanics: Theory and Experiment , 2018(3):033105, mar 2018

    2018

  13. [21]

    Energy gaps and elementary excitations for certain vbs-quantum antiferromagnets

    Stefan Knabe. Energy gaps and elementary excitations for certain vbs-quantum antiferromagnets. Journal of Statistical Physics , 52(3):627--638, Aug 1988

    1988

  14. [22]

    Finite-size criteria for spectral gaps in d-dimensional quantum spin systems, 2019

    Marius Lemm. Finite-size criteria for spectral gaps in d-dimensional quantum spin systems, 2019. arXiv preprint arXiv:1902.07141

    2019 arXiv

  15. [23]

    Gaplessness is not generic for translation-invariant spin chains

    Marius Lemm. Gaplessness is not generic for translation-invariant spin chains. Phys. Rev. B , 100:035113, Jul 2019

    2019

  16. [24]

    Spectral gaps of frustration-free spin systems with boundary, 2018

    Marius Lemm and Evgeny Mozgunov. Spectral gaps of frustration-free spin systems with boundary, 2018. arXiv preprint arXiv:1801.08915

    2018 arXiv

  17. [25]

    The aklt model on a hexagonal chain is gapped, 2019

    Marius Lemm, Anders Sandvik, and Sibin Yang. The aklt model on a hexagonal chain is gapped, 2019. arXiv preprint arXiv:1904.01043

    2019 arXiv

  18. [26]

    The spectral gap for some spin chains with discrete symmetry breaking

    Bruno Nachtergaele. The spectral gap for some spin chains with discrete symmetry breaking. Commun. Math. Phys. , 175(3):565--606, Feb 1996

    1996

  19. [27]

    Lieb-robinson bounds and the exponential clustering theorem

    Bruno Nachtergaele and Robert Sims. Lieb-robinson bounds and the exponential clustering theorem. Communications in Mathematical Physics , 265(1):119--130, Jul 2006

    2006

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.