{"id":"93c02ab3-6915-4e3d-a401-4f05dc9c2f2f","arxiv_id":"2608.04083","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"In a PT-symmetric multichannel Kondo model, exact Bethe ansatz calculations show impurity entropy becomes nonmonotonic in the zero-mode and local-moment phases, breaking generalized g-theorem irreversibility.","lead":"This paper solves exactly a non-Hermitian, PT-symmetric version of the multichannel Kondo model with two impurities and maps out four phases controlled by a non-Hermiticity parameter. It shows the impurity entropy, a measure of the defect's effective degrees of freedom, flows nonmonotonically in two phases, so a real spectrum and correct fixed-point entropies do not guarantee monotonic RG flow.","discovery_kind":"first_principles","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Zero-mode tower sums in Eqs. (17)-(21) are imported from prior work without derivation; if the tower decomposition is incomplete, the nonmonotonic impurity entropy is an artifact.","rationale":"The reader's weakest assumption identifies the same load-bearing concern: the multi-tower TBA free energy, Eqs. (17)-(21), is imported from same-author references without derivation in this Letter. This is indeed the condition that has to be true for the central claim to hold. If the tower decomposition is incomplete or the factorization into independent complex-conjugate towers is invalid, then the predicted nonmonotonic impurity entropy would be an artifact of the truncation, and the conclusion that a real spectrum plus correct endpoint defect entropies do not guarantee RG irreversibility would not follow. I checked the endpoint limits: the tower sums do reproduce the stated UV and IR entropies (for example, the T->0 limit gives 2 ln[2 cos(pi/(n+2))] per impurity), so the concern is not about internal inconsistency of the endpoints but about completeness and derivation of the intermediate temperature behavior. No internal contradiction or fitted-parameter circularity was found. The paper explicitly flags that the YSR phase is beyond the TBA scope, which is a genuine limitation but not a flaw in the central claim. The proposed concrete test would settle whether the tower sums are exact by comparing with a direct finite-size Bethe ansatz calculation or an independent re-derivation of the impurity free energy. Since the reader already conditioned the verdict on this derivation, no verdict adjustment is needed; the concern is load-bearing but appropriately captured by the CONDITIONAL verdict.","tokens_in":10348,"tokens_out":11169,"duration_ms":124580,"concrete_test":"Derive the impurity contribution to the free energy directly from the Bethe ansatz equations (2)-(3) in the zero-mode phase by enumerating all complex rapidity solutions for finite M and taking the thermodynamic limit, without importing the tower decomposition of Refs. [22,23,26]. Concretely, for n=2, alpha=3pi/4, compute S_imp(T) from Eqs. (17)-(20) and from a finite-size Bethe ansatz exact calculation for M=2,4,6 at the same couplings; if the finite-size data extrapolate to a different entropy curve, the tower sums in Eqs. (17)-(21) are missing configurations and the nonmonotonicity is not a property of the model.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central nonmonotonicity result rests entirely on the zero-mode/local-moment tower sums, Eqs. (17)-(21). These are asserted to follow from the methods of Refs. [22,23,26], but those works are Hermitian edge-impurity problems, not the PT-symmetric multichannel model studied here. The paper does not derive that the complex impurity strings in Eqs. (5)-(7) exhaust the rapidity spectrum, does not justify the integer p=floor(alpha/pi+1/2) selecting the number of higher-order strings, and does not show that the two impurities factor into independent complex-conjugate tower sums. The individual tower free energies are complex; the reality of F_imp depends on the factorization Z_imp=|Z_1|^2, which is assumed rather than proven. If additional complex rapidity solutions or tower mixing exist, the intermediate overshoots and undershoots are a truncation artifact, and the breakdown of monotonic g-flow is not established. The agreement with NRG is only qualitative, so it does not validate the exact TBA tower sums.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript studies a PT-symmetric non-Hermitian multichannel Kondo model composed of two spin-1/2 impurities coupled to n conduction channels through complex-conjugate Kondo couplings. Using a Bethe ansatz, the authors identify four impurity phases as a function of a non-Hermiticity parameter alpha: overscreened Kondo, zero mode, YSR, and local moment. A generalized thermodynamic Bethe ansatz is then used to compute the impurity free energy and the Affleck-Ludwig g-function. The central claim is that in the zero-mode phase and the local-moment phase the impurity entropy is nonmonotonic in temperature even though the spectrum is PT-unbroken and the ultraviolet and infrared impurity entropies match defect-CFT values. The paper further conjectures that RG irreversibility survives in the Kondo phase while failing in the zero-mode and local-moment phases, so that a real spectrum and correct endpoint g-values are not sufficient to guarantee monotonic g-flow.","tokens_in":10536,"tokens_out":5541,"duration_ms":60802,"significance":"If the multi-tower thermodynamic Bethe ansatz construction is correct, the paper provides a rare exact result for a non-Hermitian integrable impurity problem: the impurity entropy is obtained without fitting parameters, the infrared g-values agree with defect conformal field theory, and the calculation explains qualitatively the nonmonotonic entropy observed in non-Hermitian numerical renormalization group studies. The claimed breakdown of monotonic g-flow in a PT-unbroken phase with real spectrum and correct endpoints would be conceptually important. However, the significance is conditional on the tower-sum decomposition being exact for this model; the current manuscript does not fully establish that key premise.","major_comments":[{"comment":"The multi-tower impurity partition function is imported from Refs. [22,23,26] without a derivation. Since the nonmonotonic impurity entropy is generated entirely by the tower sums in Eqs. (17)-(21), the central claim requires either a derivation of this construction directly from the Bethe ansatz equations (3)/(39) for complex couplings or an explicit proof that the zero-energy impurity strings in Eqs. (5)-(7) exhaust the rapidity spectrum and that the towers decouple. The paper also asserts, without justification, the formula p = floor(alpha/pi + 1/2) for the number of higher-order strings. As written, the main result rests on an unproven assumption, and the reader cannot verify that the complex rapidity spectrum has been fully classified.","section":"Zero-mode phase; Eqs. (17)-(21)"},{"comment":"The reality of the total impurity free energy is guaranteed by assuming Z_imp = |Z_imp,(1)|^2, namely that the two impurities contribute independent complex-conjugate tower sums with no mixing between towers. This factorization is asserted rather than derived. If additional complex rapidity solutions or tower mixing exist, the intermediate overshoots and undershoots in Figs. 3 and 4 would be truncation artifacts rather than genuine properties of the model. The paper should either prove the factorization or provide an independent check of the tower decomposition.","section":"Eq. (19) and surrounding text"},{"comment":"The classification of impurity strings, their vanishing energies in the zero-mode phase, and the complex energies quoted for the YSR phase in Eq. (8) are load-bearing for the phase diagram and for the tower assignment used in Eqs. (17)-(21), but no derivation is given. In particular, the statement that the energies vanish identically in the thermodynamic limit and the formula determining the number of higher-order strings need to be derived from the Bethe ansatz equations, rather than stated as part of the phase classification.","section":"Zero-mode phase; Eqs. (5)-(8)"},{"comment":"The local-moment phase uses the same excitation-tower construction as the zero-mode phase, but the paper does not explain why that construction remains valid after passing through the PT-broken YSR phase. This is needed to support the cyclic-RG claim that the impurity entropy returns to 2 ln 2 through intermediate overshoots and undershoots. The reader should be told explicitly which features of the tower decomposition survive the YSR transition and why the same free-energy expressions apply.","section":"Local-moment phase; Eqs. (21)"}],"minor_comments":[{"comment":"The y-axis tick label appears as 'ln(2 + 2)' in the figure, which is likely a typesetting error for 'ln(2 + 2 phi)'; the caption introduces phi as the golden ratio, but the figure text should match the notation consistently.","section":"Fig. 2"},{"comment":"The notation F^{T1}_{(2)} and F^{T2}_{(1)} is confusing because the subscripts and superscripts are not defined explicitly. Please clarify which impurity and which tower each symbol denotes before the equations are used.","section":"Eqs. (17)-(18)"},{"comment":"The TBA hierarchy in Eq. (9) is written for the multichannel bulk, while the impurity couplings enter only later through the tower sums. It would help to state explicitly whether Eq. (9) is identical to the Hermitian multichannel TBA and where the complex couplings modify the boundary conditions or driving terms.","section":"Eq. (9)"},{"comment":"The relationship between the coupling constants in the Hamiltonian (1) and the effective parameters c and phi used throughout the paper is stated very briefly; a short explicit definition of c and phi in terms of lambda and lambda* would make the manuscript more self-contained.","section":"Introduction"}],"recommendation":"major_revision","confidential_remarks":"The paper presents an interesting and potentially important exact result, but its central nonmonotonicity claim rests on a multi-tower thermodynamic Bethe ansatz construction that is taken from three prior works by the same authors and is not independently derived here. If the authors can supply a self-contained derivation or a detailed verification of the tower decomposition for this PT-symmetric model, the paper would be a strong contribution. The qualitative agreement with NRG is encouraging but does not by itself validate the exact tower sums."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper does two genuinely new things. It writes down and solves, via Bethe ansatz, the multichannel PT-symmetric Kondo model with complex-conjugate couplings, and it reports an exact phase diagram (Kondo, zero mode, YSR, local moment) with PT-breaking in the YSR phase. The second is the conceptual claim: in the zero-mode and local-moment phases the impurity entropy is nonmonotonic in temperature even though the spectrum is PT-unbroken and the endpoint entropies match defect CFT, so real spectrum plus correct endpoints do not guarantee RG irreversibility. That is a sharp statement, and if right it matters for generalized g-theorems in non-unitary settings.\n\nWhat is solid: the Bethe equations are derived in the End Matter from the nested Bethe ansatz, and the Kondo-phase free energy gives monotonic flow to the correct IR g-value. The YSR phase is explicitly outside the paper's TBA scope, which is honest. The phase boundaries between strings are plausible, and the qualitative agreement with NRG [17,18] suggests the phenomenon is real, not a fit.\n\nThe soft spot is exactly where the reader put it. The nonmonotonicity comes from the multi-tower sums in Eqs. (17)-(21), which are not derived in this Letter. They are taken from Refs [22,23,26], which are Hermitian edge-impurity problems. Two things need to be shown: that the zero-energy impurity strings in Eqs. (5)-(7) exhaust the rapidity spectrum, and that the partition function factorizes into complex-conjugate tower sums so that F_imp is real. Neither is proved here. The integer p=floor(alpha/pi+1/2) also appears without derivation. If the tower classification is incomplete, the overshoots and undershoots are an artifact of truncation. That is a load-bearing gap, and it is the difference between an exact result and a plausible conjecture.\n\nI do not think this is fatal. The paper is honest about what is derived and what is conjectured, the setup is integrable, and the claim is falsifiable by a more complete string analysis or by direct numerical TBA. It deserves a serious referee, but the referee should ask for the tower construction to be rederived within this model, or for the code and data to ship. I would bring it to a reading group and would cite it, with a caveat.","headline":"An exact-solution claim about nonmonotonic impurity entropy in PT-symmetric multichannel Kondo, with the central caveat that the tower sums producing the nonmonotonicity are imported from prior work without derivation.","tokens_in":11075,"tokens_out":2339,"would_cite":true,"duration_ms":22353,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["72.15.Qm","11.30.Er","71.10.-w"],"model":"deepseek-v4-flash","headline":"In a PT-symmetric multichannel Kondo model, exact Bethe Ansatz shows impurity entropy can increase with cooling in some phases, breaking standard RG irreversibility.","keywords":["PT symmetry","multichannel Kondo effect","thermodynamic Bethe Ansatz","impurity entropy","Affleck-Ludwig g-function","RG irreversibility","zero-energy impurity strings","non-Hermitian quantum impurity"],"falsifier":"Solve the Bethe equations (3) numerically at finite system size without imposing the string hypothesis, keeping all complex rapidity solutions; if the impurity entropy computed from the exact root distribution is monotonic in the zero-mode or local-moment phase, the multi-tower truncation missed configurations. Conversely, reproducing the overshoot and undershoot curves with an independent nonperturbative method would confirm the claim.","tokens_in":10106,"feed_emoji":"⚛️","tokens_out":6763,"duration_ms":58597,"temperature":0.7,"pith_summary":"This paper studies a PT-symmetric multichannel Kondo model—two spin-1/2 impurities coupled to n conduction channels through complex-conjugate couplings—and solves it exactly with a thermodynamic Bethe Ansatz. The authors claim that in the zero-mode phase and in the local-moment phase the impurity entropy, which is the Affleck–Ludwig g-function, is nonmonotonic in temperature, even though the spectrum is real and the ultraviolet and infrared entropies match defect conformal field theory. The culprit is a set of zero-energy impurity strings that reorganize the Bethe-Ansatz spectrum into multiple excitation towers; in the local-moment phase the flow is cyclic and returns to the same fixed point. The result matters because it shows that a real spectrum plus correct endpoint g-values do not guarantee renormalization-group irreversibility, so any generalized g-theorem for non-Hermitian defects must impose additional conditions.","feed_headline":"Impurity entropy overshoots in PT-symmetric Kondo phases","feed_subtitle":"Exact Bethe Ansatz shows zero-energy strings make the g-function nonmonotonic; RG irreversibility fails despite correct endpoint entropies.","key_machinery":"The load-bearing object is the multi-tower impurity partition function. Zero-energy impurity string rapidities, given in Eqs. (5)–(7), appear once $\\alpha$ exceeds pi/2 and reorganize the thermodynamic Bethe-Ansatz spectrum: two towers in the zero-mode I subphase, three towers in zero-mode II and the local-moment phase. Each tower contributes a complex free energy of the form of Eqs. (17)–(21), with the two impurities' towers related by PT conjugation, so that Z_imp = |Z_imp^(1)|^2 is real and positive. The impurity entropy S_imp = ln g(T) is computed by summing these towers and differentiating the free energy; the tower reorganization is what converts a monotonically decreasing g-function into a nonmonotonic one.","core_discovery":"The central claim is the exact demonstration that monotonic impurity entropy flow—the signature of an irreversible defect RG flow—breaks in a PT-symmetric multichannel Kondo model in phases where PT symmetry is unbroken and the spectrum is real. In the Kondo phase (0<alpha<pi/2) the entropy decreases monotonically from 2 ln 2 to 2 ln[2 cos(pi/(n+2))], exactly as in defect CFT. In the zero-mode phase (pi/2<alpha<n pi/2) the same ultraviolet and infrared fixed-point entropies are reached, but the g-function develops intermediate overshoots and undershoots; in the local-moment phase (alpha>(n/2+1)pi) it returns to the UV value 2 ln 2 after overshooting and undershooting, which the authors interpret as cyclic RG flow between the same fixed point. The paper concludes that neither a real spectrum nor defect entropies consistent with defect CFT are sufficient to guarantee RG irreversibility.","pith_inferences":["If the multi-tower construction is exact, the nonmonotonicity should show up as a sign change in the impurity specific heat C_imp inside the zero-mode and local-moment phases, which the figures already display; a direct measurement of entropy in engineered gain/loss quantum-dot arrays could test this.","The mechanism likely generalizes: any integrable defect whose zero-energy bound states split the spectrum into multiple towers may exhibit nonmonotonic g-functions, independent of PT symmetry.","A generalized g-theorem for non-Hermitian defects would have to count or constrain the tower structure, not just check real spectrum and endpoint entropies.","The parameter-free predictions of the overshoot and undershoot curves can serve as a benchmark for any nonperturbative numerical treatment of non-Hermitian Kondo systems."],"forward_implications":["In the Kondo phase 0<alpha<pi/2 the g-function decreases monotonically, so RG irreversibility — and a generalized Affleck–Ludwig g-theorem — plausibly survives small departures from Hermiticity.","In both zero-mode subphases the nonmonotonic impurity entropy is tied to the zero-energy impurity strings, so the effect persists for every channel number n>=2 and grows with n.","In the local-moment phase the RG trajectory is cyclic: impurity entropy is 2 ln 2 at both the ultraviolet and infrared limits, with intermediate overshoots and undershoots.","The YSR phase spontaneously breaks PT symmetry and develops complex impurity-string energies, placing it outside the thermodynamic Bethe-Ansatz description used here.","The exact solution provides an analytic explanation for numerically observed nonmonotonic impurity entropy, identifying zero-energy impurity strings as the origin."],"supporting_citations":[{"why":"Supplies the thermodynamic Bethe Ansatz method and the single-tower impurity free energy that the Kondo-phase result extends.","marker":"[2]"},{"why":"Defines the non-Hermiticity parameter alpha and the phase diagram, including PT-broken and unbroken phases, used throughout.","marker":"[13]"},{"why":"Introduces the split-Hilbert-space thermodynamic construction with multiple towers that the present work adapts to the Kondo defect.","marker":"[22]"},{"why":"Provides the impurity-string tower method for an impurity at a superconducting edge, the direct template for Eqs. (17)–(21).","marker":"[23]"},{"why":"Generalizes the multi-tower thermodynamic Bethe Ansatz to multichannel Kondo in superconducting leads, used for the n-channel tower sums.","marker":"[26]"},{"why":"Supplies the dynamical fusion method that produces the n-channel Bethe Ansatz equations (38)–(39).","marker":"[30]"},{"why":"Defines the Affleck–Ludwig g-function and its monotonicity property, the object whose breakdown is the paper's central result.","marker":"[27]"},{"why":"Establishes the non-Hermitian Kondo effect in ultracold alkaline-earth atoms, motivating the PT-symmetric Kondo setup.","marker":"[12]"},{"why":"Reports the numerically observed nonmonotonic impurity entropy in the non-Hermitian multichannel Kondo model, which this exact solution explains.","marker":"[17]"}],"fun_headline_variants":["Entropy flow breaks monotonicity in PT-symmetric Kondo model","PT-symmetric Kondo: impurity entropy no longer monotonic","Non-monotonic impurity entropy in PT-symmetric Kondo phases","Exact solution shows entropy overshoots in PT-symmetric Kondo","g-function nonmonotonic in PT-symmetric Kondo despite real spectrum"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole argument assumes that the zero-energy impurity strings written in Eqs. (5)–(7) exhaust the new rapidity configurations, so that summing the two or three towers is the complete thermodynamics; if additional complex rapidity solutions or tower mixing exist, the predicted nonmonotonic entropy would be an artifact of the truncation.","fun_headline_variants_meta":{"raw":{"variants":["Entropy flow breaks monotonicity in PT-symmetric Kondo model","PT-symmetric Kondo: impurity entropy no longer monotonic","Non-monotonic impurity entropy in PT-symmetric Kondo phases","Exact solution shows entropy overshoots in PT-symmetric Kondo","g-function nonmonotonic in PT-symmetric Kondo despite real spectrum"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00065,"raw_usage":{"total_tokens":3076,"prompt_tokens":1136,"completion_tokens":1940,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":752,"completion_tokens_details":{"reasoning_tokens":1846}},"tokens_in":752,"tokens_out":1940,"duration_ms":14021,"temperature":1.0,"reasoning_tokens":1846,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-08T00:34:27.310400+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Solve the Bethe equations (3) numerically at finite system size without imposing the string hypothesis, keeping all complex rapidity solutions; if the impurity entropy computed from the exact root distribution is monotonic in the zero-mode or local-moment phase, the multi-tower truncation missed configurations. Conversely, reproducing the overshoot and undershoot curves with an independent nonperturbative method would confirm the claim.","supporting_citations":[{"cited_title":"Andrei, K","cited_arxiv_id":null,"evidence_quote":"Supplies the thermodynamic Bethe Ansatz method and the single-tower impurity free energy that the Kondo-phase result extends."},{"cited_title":"Kattel, A","cited_arxiv_id":null,"evidence_quote":"Defines the non-Hermiticity parameter alpha and the phase diagram, including PT-broken and unbroken phases, used throughout."},{"cited_title":"Kattel, A","cited_arxiv_id":null,"evidence_quote":"Provides the impurity-string tower method for an impurity at a superconducting edge, the direct template for Eqs. (17)–(21)."},{"cited_title":"Kattel, A","cited_arxiv_id":null,"evidence_quote":"Generalizes the multi-tower thermodynamic Bethe Ansatz to multichannel Kondo in superconducting leads, used for the n-channel tower sums."},{"cited_title":"Andrei and C","cited_arxiv_id":null,"evidence_quote":"Supplies the dynamical fusion method that produces the n-channel Bethe Ansatz equations (38)–(39)."},{"cited_title":"ground-state degeneracy","cited_arxiv_id":null,"evidence_quote":"Defines the Affleck–Ludwig g-function and its monotonicity property, the object whose breakdown is the paper's central result."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Reports the numerically observed nonmonotonic impurity entropy in the non-Hermitian multichannel Kondo model, which this exact solution explains."}],"review_version":1}