{"id":"e171f877-8ad2-4749-ae6c-6aeab472a052","arxiv_id":"2602.00486","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"In exactly solvable quadratic fermionic dissipators with partial postselection, the Liouvillian hosts a Jordan block of size n+1; many-body jump perturbations open spectral gaps scaling as z^{1/(d+1)}.","lead":"This paper finds a class of open fermion systems whose Liouvillian—the generator of dissipative quantum dynamics—has an exceptional point whose order grows with the number of sites. The work shows how the resulting algebraic decay and perturbative fractional power-law gaps could be used to detect these many-body exceptional points.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"For 1<d<n, the claimed z^{1/(d+1)} gap scaling rests on the unanalysed block J⊥,d; only n=4 numerics support it, so the universal fractional-scaling claim is not fully secured.","rationale":"The paper's central new result is the exactly solvable hybrid Liouvillian with a Jordan block of size n+1 and the associated dynamical signatures. I checked the key exact steps: the third-quantized spectrum, Eq. (31), follows from the 2×2 Jordan blocks T+(q); the Jordan chain for the mixed k-particle projectors gives L_H r0,k = -nγ r0,k + γ k r0,k-1, supporting an order-(n+1) EP; and the long-time 1/t behavior of the normalized particle number follows from that chain. These parts of the claim are solid. The load-bearing weak point is the perturbation analysis for intermediate d. The determinant in Eq. (49) only treats the largest block J0, and while the off-diagonal couplings L0^T L_d R⊥ and L⊥^T L_d R0 are argued to vanish, the internal structure of J⊥,d is not solved. Because the unperturbed Liouvillian has several Jordan blocks at the same eigenvalue λ0=-nγ, a perturbation inside J⊥,d could in principle lift them at order z or at an exponent smaller than 1/(d+1), which would dominate the spectral gap for small z. The paper's own statement that J⊥,d cannot be solved analytically, together with the numerical demonstration only at n=4, means the universal fractional scaling for 1<d<n is a conjecture supported by one small-system example. This is exactly the reader's weakest assumption. I do not see grounds to reject the paper; the concern is a missing general argument or a broader numerical test, so the conditional verdict is appropriate.","tokens_in":14725,"tokens_out":23099,"duration_ms":254646,"concrete_test":"For n=5 and n=6, construct the exact vectorized perturbed Liouvillian L'_H = L_H + zL_d on the 2^n×2^n Fock basis for d=2,...,n-1. For z=10^{-3}, 10^{-4}, ..., 10^{-8}, compute the eigenvalue with largest real part below the shifted vacuum and fit Re(λ)-λ0 to c z^α. If α deviates from 1/(d+1) for any d or n, the claim fails. Independently, compute J⊥,d in the Jordan basis and check whether its spectrum has any eigenvalue with leading order z^m for m < 1/(d+1), especially O(z) couplings among Jordan blocks at λ0=-nγ.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section III.B's fractional-scaling claim for 1<d<n depends on the complementary Jordan block J⊥,d. Eq. (47) block-diagonalizes the perturbation between J0 and J⊥, but the paper states that an analytical solution of J⊥,d is unattainable (Sec. III.B), and the z^{1/(d+1)} gap is demonstrated only for n=4 (Fig. 2(c)). At λ0=-nγ there are multiple Jordan blocks corresponding to different configurations ν with Σ_j(1-ν_j)cos(2πj/n)=0. If zJ⊥,d produces O(z) couplings among these degenerate blocks, or even a z^{1/2} splitting from a smaller Jordan block, that smaller shift would set the spectral gap and invalidate the claimed universal z^{1/(d+1)} scaling. The exact unperturbed spectrum and the order-(n+1) EP are well supported; this concern is specifically about the universal perturbative signature claimed in the abstract.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies a class of quadratic fermionic open systems governed by a hybrid Liouvillian that is not trace-preserving. Using the third-quantization formalism of Prosen, the authors solve the Liouvillian spectrum exactly for a tight-binding chain with local loss and periodic boundary conditions. They show that the spectrum is gapless, that the eigenvalue λ0 = -nγ supports a Jordan block of size n+1, and that the normalized particle number decays algebraically as 1/t at long times. They then add perturbations of the form zL_d, where L_d creates d particles, and argue that these break the exceptional point and produce a finite Liouvillian gap scaling as z^{1/(d+1)}. The unperturbed exact solution, the Jordan-chain construction, and the closed-form dynamics are the main technical results.","tokens_in":14928,"tokens_out":34328,"duration_ms":313873,"significance":"If fully established, this would be a valuable analytically solvable example of a many-body Liouvillian exceptional point whose order grows with system size, with a concrete experimentally accessible signature in the algebraic decay of the particle number. The exact spectrum in Eq. (31), the explicit generalized eigenmatrices in Eqs. (35)–(36), and the closed-form dynamics in Eqs. (38)–(39) are concrete, parameter-free results that do not rely on numerics. The perturbative analysis for d=1 (Appendix C) and d=n is also exact. However, the claimed universal fractional gap scaling for intermediate 1<d<n is not analytically established, and the paper itself states that the complementary Jordan block cannot be solved analytically. The strength of the central result is therefore uneven: the unperturbed EP physics is solid, while one of the headline perturbative claims is supported only by n=4 numerics.","major_comments":[{"comment":"The block-diagonalization of the perturbation is load-bearing, but the statement 'Similarly, we have LT_0 L_d R⊥ = 0' is not proven. Biorthogonality of l_{0,j} to R⊥ does not by itself imply Tr(l_{0,j} L_d r_{ν,k}) = 0 for a general operator L_d, since L_d changes the particle number and can move r_{ν,k} into a sector that overlaps l_{0,j}. Without this off-diagonal block vanishing, the perturbed spectrum is not simply the union of the spectra of J0+zJ0,d and J⊥+zJ⊥,d, and the fractional scaling could be modified by coupling between the blocks. Please provide a proof or an explicit computation of LT_0 L_d R⊥.","section":"Sec. III.B, Eq. (47)"},{"comment":"For 1<d<n, the universal gap scaling Δ∼z^{1/(1+d)} rests on the unanalyzed block J⊥,d. The paper states 'an analytical solution of J⊥,d is unattainable' and presents numerical evidence only for n=4 in Fig. 2(c). This is insufficient to establish the claimed universal scaling for arbitrary system size: at λ0=-nγ there are multiple Jordan blocks with different configurations, and zJ⊥,d could in principle couple these blocks and produce shifts of order z or z^{1/2} that would set the spectral gap. An analytic bound on the smallest eigenvalue shift of J⊥,d, or a systematic perturbative treatment, is needed to secure the abstract's claim of perturbation-dependent fractional power-law scalings.","section":"Sec. III.B, after Eq. (49)"}],"minor_comments":[{"comment":"The asymptotic of the normalized particle number is n0/t, not 1/t. For c_{0,n0}=1, the leading numerator and denominator terms are t^{n0-1}/(n0-1)! and t^{n0}/n0!, respectively, giving ⟨Ñ⟩_t ∼ n0/t. The abstract's statement 'approaches 1/t' should be amended accordingly, although the algebraic 1/t decay is preserved up to a constant.","section":"Eq. (40)"},{"comment":"Fig. 2(c) shows the gap scaling for n=4 only. For a claim of universal scaling in system size, a discussion of finite-size effects or at least one larger n would be helpful, even if only numerically.","section":"Sec. III.B, Fig. 2(c)"},{"comment":"There is a typo: 'Liouvllian' should be 'Liouvillian' in the introductory paragraph.","section":"Sec. I"}],"recommendation":"major_revision","confidential_remarks":"The unperturbed exact solution is strong and likely correct, but the perturbative universal scaling for intermediate d is the paper's headline claim and is not sufficiently supported. The unproven off-diagonal block LT_0 L_d R⊥ is a concrete technical gap that could affect the validity of the block-diagonalization. If the authors can prove that block vanishes and provide an analytic treatment or a rigorous bound for J⊥,d, the paper would be acceptable; otherwise, the abstract should be qualified to reflect the n=4 numerical evidence."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know. First, the unperturbed part is solid and genuinely new: the n-site chain with H = -t sum(c_j^\\dagger c_{j+1} + h.c.) and L_j = sqrt(gamma)c_j gives a Liouvillian with an n+1 Jordan block at lambda = -n gamma. The Jordan-chain construction via the mixed k-particle projectors P_k is explicit, and I checked the action: L_H P_k = -n gamma P_k + gamma(n-k+1)P_{k-1}. That chain is real, so the order-(n+1) EP is not an artifact. The resulting 1/t algebraic decay follows cleanly. That part deserves to be taken seriously. Second, the perturbative claim about universal z^{1/(d+1)} scaling of the gap for 1<d<n is not on the same footing. The paper proves the d=1 and d=n cases exactly, but for intermediate d it only analyzes J_0,d and then computes J_perp,d numerically at n=4. Since the off-diagonal block vanishes but J_perp,d itself is unanalyzed, nothing rules out a smaller Jordan block in J_perp,d producing a z^{1/m} or O(z) splitting that would set the actual gap. The stress-test note lands on this correctly. The abstract's wording—fractional power-law scalings as a signature of the EP—overreaches slightly for the intermediate regime.\n\nWhat is well done: the third-quantization formalism is used carefully, the spectral theorem is applied correctly, and the exact d=1 solution in Appendix C is a nice sanity check. The numerics are honest and not fitted; they support the claimed scaling for n=4. The citation pattern is appropriate: prior Liouvillian EP work is mostly order 2-3, and the higher-order bosonic examples are linear, not quadratic-fermionic. The paper is not circular: the spectrum and EP order are derived, not assumed.\n\nSoft spots: the unanalyzed J_perp,d in Sec. III.B is the main one. The statement that results are independent of boundary conditions is a bit quick, though for the bulk spectrum with periodic boundary conditions it is plausible. The partial-postselection interpretation is handwavy, but they cite the established hybrid-Liouvillian formalism. None of these are fatal; the central result stands.\n\nWho this is for: people working on open fermionic systems, third quantization, and exceptional-point physics. It is a constructive exactly solvable model, not a theorem paper. A serious referee should see it, but one who will push on J_perp,d. My recommendation: send to peer review, and require either an analytic argument for J_perp,d or a clear restriction of the scaling claim to the cases where it is proven. Under no circumstances desk-reject.","headline":"A genuinely solvable quadratic-fermion Liouvillian with an order-(n+1) EP, but the universal fractional-gap scaling for intermediate perturbations is only numerically supported.","tokens_in":15465,"tokens_out":1989,"would_cite":true,"duration_ms":21681,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"A quadratic fermion chain with local loss is exactly solvable and has a Liouvillian exceptional point of order n+1, so its normalized particle number decays as 1/t.","keywords":["higher-order exceptional point","Liouvillian","quadratic fermions","third quantization","dissipative dynamics","gapless spectrum","fractional power-law scaling","quantum jumps"],"falsifier":"Compute the perturbed Liouvillian spectrum for the n=4 chain with the d=2 perturbation at z=10^{-4}, 10^{-5}, and 10^{-6}; if the gap (the real part of the eigenvalue closest to -nγ) does not scale as z^{1/3}, the claimed fractional scaling fails. A d=3 run should give z^{1/4}.","tokens_in":14564,"feed_emoji":"⚛️","tokens_out":13574,"duration_ms":120504,"temperature":0.7,"pith_summary":"At issue is whether higher-order exceptional points—points where the generator of an open system's dynamics becomes defective, with eigenvalues and eigenstates coalescing—can be engineered and solved in a many-body fermionic system. The paper shows yes for a simple chain: fermions hopping between sites with each site subject to loss, studied under a partially post-selected hybrid Lindblad equation. The exact spectrum follows from a standard mapping of quadratic Lindbladians to free adjoint fermions, and one eigenvalue, -nγ, carries a Jordan block of size n+1, an exceptional point whose order grows with system size. Because this point is the quasisteady state, it governs the long-time dynamics: the spectrum is gapless and the normalized particle count relaxes algebraically as 1/t. Adding d-fermion jump perturbations lifts the degeneracy and produces Liouvillian gaps proportional to z^{1/(d+1)}, a fractional power that depends on the perturbation's particle number and can serve as a detection signature.","feed_headline":"Open fermion chain hosts an n+1-fold exceptional point","feed_subtitle":"Its highest-order degeneracy is the long-time attractor, and small perturbations open a fractional-power gap.","key_machinery":"The engine is the third-quantization mapping, which turns a quadratic Lindblad master equation into a quadratic form in 4n adjoint Majorana fermions with a 4n×4n shape matrix A. For this model, A is block-triangular and its momentum-space block is a non-diagonalizable 2×2 Jordan block at every quasimomentum—the precise source of the Liouvillian's defective nature. A counting rule for Jordan-block size, 1+Σ_j(n_j-ν_{j,k})ν_{j,k}, then gives the exceptional-point order: with n two-dimensional blocks all at the middle occupation, the largest Jordan block has size n+1. The perturbation argument uses the Jordan-form decomposition of the Liouvillian: a d-fermion jump operator acts along the Jordan","core_discovery":"The paper establishes that a simple open fermionic system can realize a Liouvillian exceptional point whose order grows with the number of sites, and that this exceptional point controls the long-time physics. For the n-site chain H=-tΣ(c†_j c_{j+1}+h.c.) with local loss L_j=√γ c_j and the hybrid Liouvillian L_Hρ=-i[H,ρ]+γΣ_j c_jρc†_j -nγρ, the eigenvalues are exactly λ_ν=2it Σ_j ν_{j,1} cos(2πj/n)-nγ with ν_{j,1}∈{0,1,2}. The eigenvalue -nγ carries a Jordan block of size n+1, so the exceptional-point order scales with system size and the spectrum is gapless. This exceptional point is the quasisteady state, so it dominates the dynamics: the normalized particle number tends to 1/t at long tim","pith_inferences":["Editorial: If the z^{1/(d+1)} scaling holds for arbitrary system size, the relaxation rate becomes increasingly sensitive to perturbation strength as n grows; this suggests a many-body analogue of exceptional-point-enhanced response, though the paper does not discuss sensing.","Editorial: The exact spectrum is derived with periodic boundary conditions; testing an open chain of a few sites would confirm whether the n+1-dimensional Jordan block survives without translational symmetry, which the paper asserts but does not demonstrate.","Editorial: The same Jordan-block counting applies to any single-band quadratic fermion chain with uniform local loss, so the construction should generalize to longer-range hopping or multiple orbitals, where the exceptional-point order may be tuned differently."],"forward_implications":["The long-time normalized particle number is universal: for every initial state it behaves as 1/t, while the unnormalized count is e^{-nγt} times a polynomial whose degree is fixed by the initial particle number; this is a direct fingerprint of the n+1-fold exceptional point.","Without perturbations the Liouvillian spectrum is gapless (all real parts equal -nγ), so the system has no unique steady state, only a quasisteady state at -nγ.","A perturbation that creates d fermions simultaneously opens a spectral gap Δ∼z^{1/(d+1)}, changing the relaxation from algebraic to exponential with a d-dependent time scale z^{-1/(d+1)}.","For d=1, adding the reverse jump c†_j makes the model exactly diagonalizable: all exceptional points disappear and the gap is γz^{1/2}; for d=n the off-diagonal coupling vanishes exactly and the gap is z^{1/(n+1)}.","Because the exceptional point sits at the quasisteady state, the fractional scaling also appears in the dynamics of the normalized particle number, making the higher-order exceptional point detectable by time-resolved particle counting."],"fun_headline_variants":["Fermion chain's Liouvillian EP order grows with system size","System-size exceptional point controls dissipative fermion dynamics","Higher-order Liouvillian EP emerges in quadratic fermions","Open fermions exhibit gapless spectrum from high-order EP"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"For intermediate perturbation orders (1<d<n), the paper assumes that the coupling between the large Jordan block and the rest of the spectrum contributes only at higher order and so the gap follows z^{1/(d+1)}; this coupling is not solved analytically and is checked numerically only for n=4.","fun_headline_variants_meta":{"raw":{"variants":["Fermion chain's Liouvillian EP order grows with system size","System-size exceptional point controls dissipative fermion dynamics","Higher-order Liouvillian EP emerges in quadratic fermions","Open fermions exhibit gapless spectrum from high-order EP"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000243,"raw_usage":{"total_tokens":1361,"prompt_tokens":735,"completion_tokens":626,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":479,"completion_tokens_details":{"reasoning_tokens":557}},"tokens_in":479,"tokens_out":626,"duration_ms":6134,"temperature":1.0,"reasoning_tokens":557,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T06:14:57.654091+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the perturbed Liouvillian spectrum for the n=4 chain with the d=2 perturbation at z=10^{-4}, 10^{-5}, and 10^{-6}; if the gap (the real part of the eigenvalue closest to -nγ) does not scale as z^{1/3}, the claimed fractional scaling fails. A d=3 run should give z^{1/4}.","supporting_citations":[],"review_version":1}