REVIEW 2 major objections 4 minor 15 references
Eta-pairing states in Hubbard models with bond-charge interactions on general graphs
T0 review · 2 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash
Pith's one-line read The paper identifies conditions under which eta-pairing states are exact eigenstates of bond-charge Hubbard models on arbitrary graphs, and constructs additional exact eigenstates of the associated Hubbard-type Hamiltonian.
desk verdict Clean central derivation of the eta-pairing condition on general graphs; the extra eigenstates in Section V are probably right but the proof is a sketch. 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 machinery has three pieces: the eta-pairing operator $\eta^\dagger=\sum_n e^{i\phi_n}a^\dagger_n b^\dagger_n$; the decomposition of each bond term into an effective hopping of strength $t_{mn}+\chi_{mn}$ plus a leftover $B_{mn2}$ satisfying $[[B_{mn2},\eta^\dagger],\eta^\dagger]=0$, which makes $B_{mn2}$ act as zero on eta-pairing states; and the phase condition $f_{mn}=(t_{mn}+\chi_{mn})e^{i\phi_n}+(t_{mn}+\chi_{mn})^* e^{i\phi_m}=0$. The condition $f_{mn}=0$ is equivalent to the loop product Eq. (40), and it is what turns $\eta^\dagger$ into an eigenoperator of $H_m$.
What would settle it
Compute, on a small graph containing two odd cycles that share a site, the overlap of $(T_b)^M b^\dagger_n|0\rangle$ with $b^\dagger_n|0\rangle$ for an odd integer $M>1$, equivalently the norm of $H_m(T_b)^M\eta^\dagger|0\rangle$; a nonzero value would show the odd-loop cancellation lemma fails and $|\psi_{NM}\rangle$ is not an eigenstate of $H_m$ for that graph.
Extended reading notes
Core claim
On the paper's own terms, the central discovery is an equivalence and a new family of eigenstates. The equivalence says that $(\eta^\dagger)^N|0\rangle$ is an eigenstate of $H=\sum_{\langle m,n\rangle}(T_{mn}+B_{mn})+U\sum_n a^\dagger_n a_n b^\dagger_n b_n$ if and only if it is an eigenstate of $H_m$, because the leftover part $B_{mn2}$ annihilates every eta-pairing state. For $H_m$, the commutator $[H_m,\eta^\dagger]=U\eta^\dagger$ holds exactly when the loop condition Eq. (40) is satisfied, and then $H_m(\eta^\dagger)^N|0\rangle=NU(\eta^\dagger)^N|0\rangle$. The new family is $|\psi_{NM}\rangle=(\eta^\dagger)^{N-1}(T_b)^M\eta^\dagger|0\rangle$ with odd $M$, claimed to be eigenstates of $H_m$ with energy $(N-1)U$, distinct from the eta-pairing states.
Load-bearing premise
The argument for the new eigenstates depends on the unproved claim that a spin-down electron that hops an odd number of times has zero amplitude to return to its starting site, because contributions from traversing an odd loop in opposite directions cancel; this is what guarantees no double occupation in $(T_b)^M\eta^\dagger|0\rangle$.
Editorial extensions
If this is right
- For any specified phases in $\eta^\dagger$, there exists a bond-charge Hubbard model, with the phase of $t_{mn}+\chi_{mn}$ chosen as in Eq. (37), for which $(\eta^\dagger)^N|0\rangle$ is an exact eigenstate.
- For any specified Hamiltonian whose effective hopping phases satisfy Eq. (40) around every loop, one can choose $\eta^\dagger$ so that $(\eta^\dagger)^N|0\rangle$ is an exact eigenstate of both $H_m$ and $H$ with energy $NU$.
- When Eq. (40) holds, $|\psi_{NM}\rangle=(\eta^\dagger)^{N-1}(T_b)^M\eta^\dagger|0\rangle$ with odd $M$ are exact eigenstates of the Hubbard-type Hamiltonian $H_m$ with energy $(N-1)U$.
- If $t_{mn}+\chi_{mn}=0$ for every bond, the condition imposes no restriction on the phases of $\eta^\dagger$, so eta-pairing states with arbitrary phases are eigenstates.
Reading between the lines
- A direct numerical check of the new eigenstates would be to compute the norm of $H_m(T_b)^M\eta^\dagger|0\rangle$ on a small graph with two odd cycles sharing a site; a nonzero value would falsify the odd-loop cancellation used in Section V.
- The same double-commutator structure suggests that other interaction terms whose double commutator with $\eta^\dagger$ vanishes could be added to $H$ without destroying the eta-pairing states, generalizing the bond-charge construction beyond $B_{mn}$.
- If the additional eigenstates are scar states, they should appear as low-entanglement outlier eigenstates in exact diagonalization spectra of $H_m$ on non-bipartite graphs; this is a testable prediction the paper does not make.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper studies Hubbard models with bond-charge interactions on arbitrary graphs. It decomposes the bond-charge term into a modified hopping term plus a residual interaction B_mn2, and shows that B_mn2 annihilates every eta-pairing state (eta^dag)^N|0>. It follows that the eta-pairing state is an eigenstate of the original Hamiltonian H exactly when it is an eigenstate of the modified Hubbard Hamiltonian H_m of Eq. (24). The paper then derives the commutator [H_m, eta^dag], obtains the phase condition fmn=0 in Eq. (33), and recasts it as the loop condition Eq. (40). Under this condition, (eta^dag)^N|0> is an exact eigenstate of H and H_m. Section V further claims new eigenstates of H_m of the form |psi_NM> = (eta^dag)^{N-1}(T_b)^M eta^dag|0> for odd M, based on the assertion that H_m(T_b)^M eta^dag|0> = 0.
Significance. If the results are fully established, the paper provides a clean, general-graph criterion for eta-pairing eigenstates in bond-charge Hubbard models, unifying and extending earlier results for hypercubic and triangular lattices. The central derivation is self-contained and does not rely on fitted parameters or on assuming the desired conclusion. The loop condition Eq. (40) is an elegant and falsifiable characterization. The claimed additional eigenstates in Eq. (45) are interesting as potential quantum many-body scar states for H_m, but their proof currently rests on an unproved combinatorial lemma, so the significance of Section V is conditional on that lemma being rigorously supplied.
major comments (2)
- [Section V, after Eq. (47)] The proof of Eq. (44) for odd M is incomplete. The manuscript asserts that for an odd number of hops a spin-down electron has zero amplitude to return to its starting site because any odd return must go around an odd loop and the two opposite directions cancel by Eq. (40). This does not constitute a proof for arbitrary closed walks: a walk may traverse several odd cycles, repeat edges, or backtrack, and no involution pairing of walks is exhibited. Since Eq. (44) is the load-bearing step for the new eigenstates in Eq. (45), this lemma must be proved. I verified independently that the lemma is true: under Eq. (36), the matrix D^{-1}(t+chi)D with D=diag(e^{i phi_n/2}) is i times a real skew-symmetric matrix, so its odd powers have zero diagonal, implying (T_b)^M b^dag_n|0> has no component at site n for odd M. The manuscript should include this or an equivalent rigorous argument.
- [Section II, Eq. (33)] The text states that Eq. (33) is 'the condition' for (eta^dag)^N|0> to be an eigenstate of H_m and H, but the derivation only establishes sufficiency, and necessity is false in general. For example, on a two-site graph with N=2, the state (eta^dag)^2|0> is the fully filled state and is annihilated by every hopping term for any choice of t_mn+chi_mn, so it is an eigenstate of H_m and H even when fmn is nonzero. The abstract and Section II should either explicitly state that Eq. (33) is a sufficient condition under which the eta-pairing method applies, or add the missing hypotheses and proof needed for a true necessary-and-sufficient statement.
minor comments (4)
- [Section II] The word 'communicator' should be 'commutator' in the sentence 'it is needed to calculate the communicator [Hm, eta^dag]'.
- [Section V] The text says 'for odd integer M' but the argument requires M to be a positive odd integer; M=0 would not satisfy Eq. (44). This should be stated explicitly.
- [Section III] In the square-lattice example, the statement that the eta-pairing state is an eigenstate when nearest-neighbour hoppings are real and next-nearest-neighbour hoppings are pure imaginary should specify that this follows for the bipartition example in Eq. (39) when the next-nearest-neighbour bonds connect sites within the same sublattice; otherwise the claim is ambiguous.
- [Eq. (21)] The derivation of Eq. (21) uses the relation [B_mn2, eta^dag](eta^dag)^N|0> = (eta^dag)^N[B_mn2, eta^dag]|0>, which is valid because of Eq. (8), but this step is worth spelling out in one sentence for readability.
Circularity Check
No significant circularity: Eq. (40) is derived from commutator algebra rather than assumed, the only self-citation [4] is introductory background, and the Section V odd-loop lemma is an unproved-but-independently-true assertion that is a rigor gap, not a circular step.
full rationale
The paper's central claim, that Eq. (40) is the condition for eta-pairing eigenstates, is obtained by direct computation rather than assumed. The Hamiltonian is decomposed as H = Hm + Sum Bmn2 (Eq. 25); the commutator chain [[Bmn2, eta-dagger], eta-dagger] = 0 (Eqs. 8-20) proves Bmn2 (eta-dagger)^(N+1)|0> = 0 (Eq. 23), so the eigenstate condition for H is identical to that for Hm. The Hm condition is computed explicitly as fmn = 0 (Eq. 33), equivalently Eq. (36), and Eq. (40) follows by multiplying Eq. (36) around a loop, with the converse given by a standard phase-consistency assignment. No step defines the conclusion in terms of itself. The only self-citation is [4] (Ye and Lin 2018), used in the Introduction solely as background ('There are also some efforts to find new exact eigenstates of the Hubbard model using other methods [4]') and never invoked in any derivation, so it is not load-bearing. The one genuine weakness is the omitted proof of the odd-loop cancellation lemma in Section V (after Eq. 47): the claim that an odd closed walk has zero amplitude to return to its starting site because the two directions around an odd loop cancel is asserted, not proven for walks with repeated edges or several odd cycles, leaving Eq. (44) resting on an unproved (though true) lemma. This is an incompleteness in the derivation, not circularity, because the lemma does not assume the conclusion of Eq. (44) and is independently verifiable (with phases chosen per Eq. (36), the single-particle hopping matrix is gauge-equivalent to a skew-symmetric matrix, whose odd powers have zero diagonal). No fitted parameter is relabeled as a prediction; all conditions are exact algebraic or graph-theoretic constraints on the Hamiltonian parameters, so the derivation is self-contained.
Assumptions & free parameters
assumptions (3)
- standard math Fermionic canonical anticommutation relations for a_n, a†_n, b_n, b†_n
- domain assumption The graph Hamiltonian is Hermitian, with t_nm = t*_mn and χ_nm = χ*_mn
- domain assumption The integer N is at most the number of sites, so (η†)^N |0⟩ is nonzero
Cite this review
Pith. "Pith review of Eta-pairing states in Hubbard models with bond-charge interactions on general graphs." pith.science (2026). https://pith.science/paper/3TQC4UXD
@misc{pith2026250601553,
author = {Pith},
title = {Pith review of: Eta-pairing states in Hubbard models with bond-charge interactions on general graphs},
year = {2026},
howpublished = {\url{https://pith.science/paper/3TQC4UXD}},
note = {Machine review of arXiv:2506.01553}
}
abstract
We investigate Hubbard models with bond-charge interactions on general graphs. For a Hamiltonian \(H\) of such a model, we provide the condition on its parameters under which the \(\eta\)-pairing method can be employed to construct its exact eigenstates. We arrive at this condition by finding that the requirement for the \(\eta\)-pairing state \((\eta^\dagger)^N |0\rangle\) to be an eigenstate of \(H\) is identical to the requirement for it to be an eigenstate of a Hubbard-type Hamiltonian \(H_m\). When the condition for \((\eta^\dagger)^N |0\rangle\) to be an eigenstate of the Hubbard-type Hamiltonian \(H_m\) is satisfied, we demonstrate that there are additional states, distinct from \((\eta^\dagger)^N |0\rangle\), which are also exact eigenstates of \(H_m\). Our results enhance the understanding of Hubbard models on general graphs, both with and without bond-charge interactions.
Reference graph
Works this paper leans on
- [4]
-
[1]
J. C. Hubbard, Proceedings of the Royal Society A 276, 238 (1963)
work page 1963
-
[2]
D. P. Arovas, E. Berg, S. A. Kivelson, and S. Raghu, An- nual Review of Condensed Matter Physics13, 239 (2022)
2022
-
[3]
C. N. Yang, Phys. Rev. Lett. 63, 2144 (1989)
1989
- [5]
-
[6]
S. Moudgalya, N. Regnault, and B. A. Bernevig, Phys. Rev. B 102, 085140 (2020)
work page 2020
-
[7]
X. Z. Zhang and Z. Song, Phys. Rev. B 103, 235153 (2021)
work page 2021
- [8]
Show all 15 references
-
[9]
Nakagawa, H
M. Nakagawa, H. Katsura, and M. Ueda, Phys. Rev. Res. 6, 043259 (2024)
2024
-
[10]
J. E. Hirsch and F. Marsiglio, Phys. Rev. B 39, 11515 (1989)
1989
-
[11]
J. M. Deutsch, Phys. Rev. A 43, 2046 (1991)
1991
-
[12]
Srednicki, Phys
M. Srednicki, Phys. Rev. E 50, 888 (1994)
1994
-
[13]
D. K. Mark and O. I. Motrunich, Phys. Rev. B 102, 075132 (2020)
2020
-
[14]
Arrachea and A
L. Arrachea and A. A. Aligia, Phys. Rev. Lett. 73, 2240 (1994)
1994
-
[15]
Schadschneider, Phys
A. Schadschneider, Phys. Rev. B 51, 10386 (1995)
1995
Reviewed August 7, 2026 · model on record in the stance chip above.
Discussion (0). Sign in to comment.