{"id":"3cdae1eb-87d3-424d-91df-b77a0ba03f8b","arxiv_id":"hep-th/0112045","paper_version":4,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"In five-dimensional Einstein-Gauss-Bonnet gravity, charged black hole entropy can be negative for some parameter ranges, which the authors propose signals a new phase transition between de Sitter and anti-de Sitter black holes.","lead":"Charged black holes in five-dimensional Einstein-Gauss-Bonnet gravity are shown to have a Hawking-Page phase diagram with a critical line rather than a point, and for some parameter ranges their entropy comes out negative. The authors propose that negative entropy signals a new instability in which de Sitter and anti-de Sitter black holes transmute into one another.","discovery_kind":"extension","skeptic_critique":{"model":"selfhosted/deepseek-v4.1-flash","headline":"The negative-entropy result (Eq. 170) is obtained only from S = ∫dE/T with E taken as the counterterm mass; it is never cross-checked against the free energy of Eq. 58 or the Wald entropy, and it carries an unremarked pole at ε = 1/4.","rationale":"The paper is a competent calculation: the GB charged AdS solution, T_H, F, and phase diagram are standard and internally consistent, and the speculative interpretation is acknowledged. The one place where the headline could actually fail is the entropy formula (170), which is the sole basis for the negative-entropy claim and is derived by a single route (∫dE/T with E = counterterm mass) without the free-energy or Wald cross-check that the paper itself makes available. The reader's S0 concern is real but secondary: it only affects small r_H (Eq. 173), not the large-r_H region 1/12 < ε < 1/4 that carries the headline. The ε = 1/4 pole is an unremarked feature that should be explained. None of this shows the paper is wrong; it identifies the check that would settle it. Verdict unchanged: CONDITIONAL is appropriate.","tokens_in":28429,"tokens_out":8315,"duration_ms":140287,"concrete_test":"Independently compute the entropy of the 5d GB black hole two ways and compare with Eq. (170): (a) S = -∂F/∂T_H from the free energy in Eq. (58), holding the appropriate charge/area fixed; (b) the Wald entropy S = -2π∫(∂L/∂R_μνρσ)ε_μν ε_ρσ √h for the GB Lagrangian. If either disagrees with (170) in the coefficient of the ε-corrections (in particular the (1-12ε)/(1-4ε) prefactor and the 24εkπr_H term), the negative-entropy region is not established.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The headline (abstract:s7, §8) is that S can be negative for 1/12 < ε < 1/4. The entropy is computed in Eqs. (169)-(170) as S = ∫dE/T with E the counterterm mass (167) and µ(r_H) from (168). Two things are assumed: (i) the counterterm mass equals the thermodynamic energy, and (ii) S0 = 0. For large r_H the sign of (170) is S0-independent, since the leading term is (V3/κ²)((1-12ε)/(1-4ε))4πr_H³; so the reader's S0 worry does not by itself remove the large-black-hole negative region. The sharper issue is (i). The paper has an independent handle it does not use: the free energy F(r_H, T_H) of Eq. (58), from which S = -∂F/∂T should follow. For c = 0 the two routes agree (Eqs. 150 and 156), which is reassuring, but no such check is performed for the GB case. Moreover (170) has a pole at ε = 1/4 (traceable to ∂µ/∂r_H in (168)), exactly at the upper edge of the claimed negative region, and the paper does not discuss it; a divergent entropy at a finite coupling is a warning that the mass-energy identification or the parameterization breaks down there. If (170) is not the Wald entropy, the 'critical line' of negative entropy is an artifact of the mass-energy identification rather than a property of the black hole.","agreement_with_reader":"partial"},"referee_report":{"model":"selfhosted/deepseek-v4.1-flash","summary":"The manuscript studies black-hole thermodynamics in five-dimensional Einstein-Gauss-Bonnet gravity with a Maxwell field. For the charged Schwarzschild-AdS branch it derives the solution (Eq. 37), the horizon equation (Eq. 38), the Hawking temperature (Eq. 39), the regularized action and free energy (Eqs. 45-49), and draws Hawking-Page phase diagrams with a critical line in the Gauss-Bonnet coefficient ε and the charge Q (Figs. 1-2). It then constructs surface counterterms for general R² gravity on SAdS and SdS backgrounds (Eqs. 130-143), obtains the conserved mass, and uses S = ∫dE/T with that mass to argue that the entropy can be negative or zero for a range of ε (Eq. 170). The authors speculate that the negative-entropy region signals a new instability with transitions between SdS and SAdS black holes. A section on dS/CFT brane-gravity localization is also included.","tokens_in":28801,"tokens_out":15320,"duration_ms":305953,"significance":"If the negative-entropy result holds, it is a striking statement about higher-derivative gravity: thermodynamic entropy need not be positive, which bears on the Wald entropy, on dS/CFT entropy bounds, and on the stability of Gauss-Bonnet black holes. The explicit counterterm construction of Eqs. (130)-(143) and the phase-diagram analysis of Figs. 1-2 are solid technical contributions that stand on their own. The paper is appropriately explicit that the SdS/SAdS transition scenario is speculative, and it does not overclaim a dynamical mechanism.","major_comments":[{"comment":"The negative-entropy result is obtained only from S = ∫dE/T with E taken as the counterterm mass (167) and µ(r_H) from (168). The paper has an independent handle it does not use: the free energy F(r_H, T_H) of Eq. (58), from which S = -∂F/∂T|_Q should follow using the constraint (57). For c = 0 the two routes agree (Eqs. 150 and 156), which is reassuring, but no such check is performed for the Gauss-Bonnet case. Since the sign of S is the headline claim, this check is load-bearing. A second, related check is the Wald entropy: for higher-derivative gravity the thermodynamic entropy and the Wald entropy need not coincide, and the paper should state which it regards as physical and whether (170) agrees with the Wald formula in the appropriate limit.","section":"§8, Eq. (170)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The prefactor (1-12ε)/(1-4ε) has a pole at ε = 1/4, exactly the upper edge of the claimed negative region 1/12 < ε < 1/4. At ε = 1/4, Eq. (57) shows that the coefficient (4c/l² - 1/κ²) vanishes, so µ is no longer determined by r_H and T_H and the parametrization degenerates. The paper does not discuss this; the 'critical line' boundary at ε = 1/4 may be an artifact of the chart rather than a physical boundary. The domain of validity of (170) should be stated and the ε → 1/4 limit analyzed.","section":"§8, Eq. (170)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The Gauss-Bonnet-specific negative entropy (Eq. 170) and the SdS↔SAdS transition speculation (based on Eqs. 163-166, which are for the c = 0 R² case) are not connected. The paper should state whether the Gauss-Bonnet entropy also exhibits the complementary sign behavior (SAdS negative while SdS positive, and vice versa) that the transition scenario requires, or restrict the speculation to the c = 0 case.","section":"§8","authors_acknowledge":null,"authors_acknowledge_p":null}],"minor_comments":[{"comment":"Eq. (98) gives S = π²r³/(4G), but the thermodynamic relation (99) with the mass (97) and the c = 0 limit of Eq. (150) both require S = π²r³/(2G); the factor of two should be corrected or the convention stated.","section":"§5, Eq. (98)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The abstract says the surface counterterms are used to get the conserved mass, free energy and entropy, but the free energy of §2 (Eq. 58) comes from the action regularization, not from the counterterms; the sentence should be rephrased.","section":"Abstract","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The claim that brane-gravity localization in dS is 'demonstrated' overstates the argument, which is an analogy with [18] under l² → -l² and does not include a perturbation analysis or an explicit counterterm evaluation.","section":"§4","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The small-black-hole negative region r_H² < -12ε is stated only for S0 = 0; unlike the large-r_H region, its sign is not S0-independent and this should be flagged explicitly.","section":"§8, Eq. (173)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"Eq. (170) is specialized to k = 2 (spherical horizon) through Eq. (168); this should be stated where the formula is presented.","section":"§8, Eq. (170)","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"The assertion that there are no contributions from B̃_μν and C̃_μνρσ is stated without demonstration; a line of justification would help the reader.","section":"§7","authors_acknowledge":null,"authors_acknowledge_p":null},{"comment":"There are numerous typographical errors, including 'of of' and 'maybe localized' in the abstract, 'appearence', 'Schwarzshild', 'balck hole', 'coodinate', 'Acknoweledgements', 'Spardlin' for Spradlin, and 'Iver' for Iyer.","section":"General","authors_acknowledge":null,"authors_acknowledge_p":null}],"recommendation":"major_revision","confidential_remarks":"The paper is a broad 2002 technical work whose headline (negative entropy) is interesting but rests on an unperformed cross-check against the free energy of Eq. (58) and against the Wald entropy. The counterterm construction is a genuine contribution and the phase-diagram analysis is sound. I would ask the authors to perform the -∂F/∂T check and to discuss the ε = 1/4 degeneration before the negative-entropy claim is published as stated."},"author_rebuttal":null,"desk_editor":{"model":"selfhosted/deepseek-v4.1-flash","letter":"Worth reading for the machinery, not for the headline. The solid part is sections 2 and 7. The charged GB-AdS solution, the Hawking temperature, and the free energy are done with standard algebra and check out; the phase diagram with a critical line rather than a critical point (Eqs. 53, 59) is a real extension of Cai's uncharged case, and the counterterm construction for general R2 gravity on SAdS/SdS backgrounds (Eq. 141) reproduces the Einstein results in the right limit (Eqs. 142-143). That mass formula is the piece I would actually cite. The negative-entropy claim is where I would push. The authors are upfront that it is speculation, and negative Wald entropy in higher derivative gravity is a real thing, so I do not think the idea is wrong. But the derivation has a gap the paper does not close. Eq. 170 comes from S = integral dE/T with E the counterterm mass and S0 = 0. The paper has an independent handle it never uses: the free energy F(r_H, T_H) of Eq. 58, from which S = -dF/dT should follow. For c = 0 the two routes agree (150 vs 156), which is reassuring, but no such check is done for GB. Until it is, we do not know whether (170) is the Wald entropy or an artifact of identifying counterterm mass with thermodynamic energy. The sharper warning is the pole at epsilon = 1/4 in (170), traceable to the partial derivative of mu with respect to r_H in (168). That sits exactly at the upper edge of the claimed negative region 1/12 < epsilon < 1/4. A divergent entropy at finite coupling usually means the parameterization or the mass-energy identification has broken down, and the paper does not mention it. Smaller things: the dS/CFT localization argument is a heuristic transposition of Hawking-Hertog-Reall, and the paper says a direct proof is still owed. The higher-derivative Maxwell section stops after Eq. 68 with 'extremely complicated' - fine as a flag, not a result. The reader's S0 worry is weaker than it looks: for large r_H the leading term in (170) is S0-independent, so the large-black-hole negative region survives any shift of the zero. Send it out. The calculation deserves referee time; the entropy section needs the free-energy cross-check before the phase-transition language is taken seriously.","headline":"Competent Gauss-Bonnet thermodynamics with a genuinely useful counterterm mass formula, but the negative-entropy headline rests on an unclosed gap between the counterterm mass and the free energy.","tokens_in":29484,"tokens_out":1679,"would_cite":true,"duration_ms":26583,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"pith_extraction":{"msc":[],"pacs":["98.80.Hw","04.50.+h","11.10.Kk","11.10.Wx"],"model":"selfhosted/deepseek-v4.1-flash","headline":"In five-dimensional Einstein-Gauss-Bonnet gravity, the thermodynamic entropy of Schwarzschild-de Sitter and Schwarzschild-anti-de Sitter black holes can come out negative or zero for a range of the higher-derivative coupling, and the paper,","keywords":["Einstein-Gauss-Bonnet gravity","black hole thermodynamics","negative entropy","Hawking-Page phase transition","Schwarzschild-de Sitter black hole","surface counterterms","de Sitter/conformal field theory correspondence","charged black holes in anti-de Sitter space"],"falsifier":"Recompute the same Gauss-Bonnet black hole entropy with the Noether-charge (Wald) formula, which does not require integrating $dE/T$ or fixing $S_0$, and check whether the negative-entropy window $1/12 <  < 1/4$ survives; if the sign is scheme-dependent there, the claimed instability has no physical content.","tokens_in":29717,"feed_emoji":"🕳️","tokens_out":2649,"duration_ms":47629,"temperature":0.7,"pith_summary":"The paper asks what happens to black hole thermodynamics when the gravitational action carries a Gauss-Bonnet term, the curvature-squared combination that keeps the field equations second order and arises as the next-to-leading correction in string effective actions. Working in five dimensions with an electromagnetic field, the authors solve for charged Schwarzschild-anti-de Sitter black holes, compute the Hawking temperature, the regularized action and the free energy, and draw the Hawking-Page phase diagram between the black hole and pure anti-de Sitter space. Because the Gauss-Bonnet coefficient and the electric charge are both free, the boundary between the two phases is a critical line in the parameter plane rather than a single critical point. The authors then derive the surface counterterms needed to make the action finite on asymptotically de Sitter and anti-de Sitter backgrounds in higher-derivative gravity, use them to obtain a conserved black hole mass, and feed that mass into $S = 8$ to get the entropy. In a range of couplings the resulting entropy is negative or zero, and the authors propose that this marks a new kind of instability in which a black hole with negative entropy in one asymptotic geometry gives way to one with positive entropy in the other.","feed_headline":"Higher-derivative gravity can give black holes negative entropy","feed_subtitle":"Integrating dE/T with a counterterm mass flips the entropy sign for some Gauss-Bonnet couplings, hinting at a dS/AdS transition.","key_machinery":"Two pieces of machinery carry the argument. The first is the Gauss-Bonnet solution itself: with $a = c$ and $b = -4c$ in the curvature-squared action, the metric function $e^{2} = e^{-2}$ solves a quadratic whose roots give the horizon radii, the Hawking temperature $4 T_H = (e^{2})'$ at $r = r_H$, and a free energy regularized by subtracting the $= Q = 0$ background. The second is the surface-counterterm construction: auxiliary fields $A$, $B_{}$, $C_{}$ linearize the $R^2$ action, a Dirichlet-type boundary condition is imposed on the scalar contractions, and the resulting boundary term plus a Gibbons-Hawking term and two counterterms with coefficients $1$ and $2$ yields a finite stress tensor and a conserved mass $M = (3l^2/16) V_3 (1/^2 - 40a/l^2 - 8b/l^2 - 4c/l^2)(k^2 + 16 l^2)$. That mass is what the entropy integral consumes, so the counterterm coefficients set the sign of the entropy.","core_discovery":"The central claim is that in higher-derivative gravity the entropy of a Schwarzschild-de Sitter or Schwarzschild-anti-de Sitter black hole need not be positive. Using surface counterterms derived for general $R^2$ gravity on these backgrounds, the authors obtain a conserved mass $M$, identify it with the thermodynamic energy $E$, and integrate $dS = dE/T$ to get the entropy. For the Gauss-Bonnet case the result is $S = (V_3/8)((1-12)/(1-4)) (4 r_H^3 + 24 k r_H) + S_0$ with $= c^2/l^2$, which turns negative for $1/12 <  < 1/4$ and, when $S_0 = 0$, for small black holes with $r_H^2 < -12$ in the $ < 0$ branch. Crucially, the sign flips the other way between the two backgrounds: the parameter range that makes the anti-de Sitter entropy negative makes the de Sitter entropy positive, and vice versa. The authors read this as a possible new type of phase transition, or 'phase transmutation', between de Sitter and anti-de Sitter black holes, passing through a state of zero mass, zero free energy and zero entropy.","pith_inferences":["The negative-entropy region is only as robust as the choice $S_0 = 0$; a nonzero integration constant would shift the zero of entropy and could remove the negative region for small black holes while leaving the large-black-hole statement intact.","An independent check would be to compute the Gauss-Bonnet black hole entropy by the Noether-charge (Wald) method, which does not rely on integrating $dE/T$, and compare signs; agreement would turn the observation from a convention-dependent statement into a physical one.","If the proposed de Sitter/anti-de Sitter transmutation is real, it should show up as a discontinuity or a runaway in the finite-temperature effective potential of the dual boundary theory, which is in principle a computable quantity rather than a speculation.","Because the effect is driven by the curvature-squared term competing with the Einstein term, it should scale with the ratio of the Gauss-Bonnet coupling to the cosmological scale, suggesting a sharp threshold rather than a gradual onset that a numerical scan of the phase diagram could confirm."],"forward_implications":["The Hawking-Page diagram for charged anti-de Sitter black holes in Gauss-Bonnet gravity acquires a critical line, not a critical point, because the Gauss-Bonnet coefficient and the charge can be traded against each other.","Surface counterterms exist for higher-derivative gravity on de Sitter as well as anti-de Sitter backgrounds, so a conserved black hole mass and a finite free energy can be defined in both cases.","The entropy sign is background-dependent: parameters that make the anti-de Sitter black hole entropy negative leave the de Sitter entropy positive, so a single theory can host a stable phase in one asymptotic geometry and an unstable one in the other.","At the critical coupling where the counterterm mass vanishes, black holes of any horizon radius become massless, which the authors suggest would allow black holes to be produced copiously.","Bounds on de Sitter entropy proposed for Einstein gravity, such as the conjecture that de Sitter entropy bounds asymptotically de Sitter entropy, cannot hold in higher-derivative gravity once the entropy can be zero or negative."],"supporting_citations":[{"why":"Supplies the surface counterterm method in higher-derivative gravity and the replacement $1/^2 1/^2 + 40a/l^2 + 8b/l^2 + 4c/l^2$ that fixes the conserved mass.","marker":"[6]"},{"why":"Provides the auxiliary-field linearization of the $R^2$ action and the boundary terms used to make the variational principle well defined.","marker":"[23]"},{"why":"Supplies the counterterm stress-tensor and mass computation in asymptotically anti-de Sitter space that the higher-derivative generalization is checked against.","marker":"[24]"},{"why":"Supplies the de Sitter energy and the Nariai-state interpretation used to conjecture the conserved mass in the de Sitter case.","marker":"[19]"},{"why":"Provides the Gauss-Bonnet black hole solution that the charged solution reduces to when the charge is set to zero.","marker":"[9]"},{"why":"Defines the Hawking-Page phase transition that the critical-line analysis extends to charged Gauss-Bonnet black holes.","marker":"[7]"},{"why":"Supplies the brane-gravity localization argument by gluing two bulk balls on a boundary, which the paper transposes to de Sitter space.","marker":"[18]"},{"why":"Provides the de Sitter/conformal field theory correspondence that motivates the de Sitter branch of the thermodynamics.","marker":"[2]"}],"fun_headline_variants":["Gauss-Bonnet black holes can have negative entropy","Black hole entropy flips sign between dS and AdS","Higher-derivative terms can push black hole entropy below zero","Counterterm mass yields negative entropy for SdS and SAdS","Negative entropy hints at a dS/AdS black hole transition"],"cache_read_input_tokens":0,"weakest_assumption_plain":"The entropy is obtained by integrating $dE/T$ with the integration constant set to zero and with the counterterm-derived conserved mass identified as the thermodynamic energy, so the reported sign of the entropy rests on that normalization choice and on that identification.","fun_headline_variants_meta":{"raw":{"variants":["Gauss-Bonnet black holes can have negative entropy","Black hole entropy flips sign between dS and AdS","Higher-derivative terms can push black hole entropy below zero","Counterterm mass yields negative entropy for SdS and SAdS","Negative entropy hints at a dS/AdS black hole transition"]},"model":"selfhosted/deepseek-v4.1-flash","effort":"low","attempts":1,"cost_usd":0.0,"raw_usage":{"total_tokens":1632,"prompt_tokens":1043,"completion_tokens":589,"prompt_tokens_details":null,"completion_tokens_details":{"reasoning_tokens":515}},"tokens_in":1043,"tokens_out":589,"duration_ms":11231,"image_pages":[],"temperature":1.0,"reasoning_budget":50,"reasoning_tokens":515,"transport_provider":"selfhosted","cache_read_input_tokens":0,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-10-29T17:42:58.290674+00:00","model_set":{"reader":"selfhosted/deepseek-v4.1-flash"},"falsifier":"Recompute the same Gauss-Bonnet black hole entropy with the Noether-charge (Wald) formula, which does not require integrating $dE/T$ or fixing $S_0$, and check whether the negative-entropy window $1/12 <  < 1/4$ survives; if the sign is scheme-dependent there, the claimed instability has no physical content.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines the Hawking-Page phase transition that the critical-line analysis extends to charged Gauss-Bonnet black holes."},{"cited_title":"Holographic entropy and brane FRW-dynamics from AdS black hole in d5 higher derivative gravity","cited_arxiv_id":"hep-th/0105117","evidence_quote":"Supplies the surface counterterm method in higher-derivative gravity and the replacement $1/^2 1/^2 + 40a/l^2 + 8b/l^2 + 4c/l^2$ that fixes the conserved mass."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the auxiliary-field linearization of the $R^2$ action and the boundary terms used to make the variational principle well defined."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the counterterm stress-tensor and mass computation in asymptotically anti-de Sitter space that the higher-derivative generalization is checked against."},{"cited_title":"Mass, Entropy and Holography in Asymptotically de Sitter Spaces","cited_arxiv_id":"hep-th/0110108","evidence_quote":"Supplies the de Sitter energy and the Nariai-state interpretation used to conjecture the conserved mass in the de Sitter case."},{"cited_title":"Gauss-Bonnet Black Holes in AdS Spaces","cited_arxiv_id":"hep-th/0109133","evidence_quote":"Provides the Gauss-Bonnet black hole solution that the charged solution reduces to when the charge is set to zero."},{"cited_title":"Brane New World","cited_arxiv_id":"hep-th/0003052","evidence_quote":"Supplies the brane-gravity localization argument by gluing two bulk balls on a boundary, which the paper transposes to de Sitter space."},{"cited_title":"The dS/CFT Correspondence","cited_arxiv_id":"hep-th/0106113","evidence_quote":"Provides the de Sitter/conformal field theory correspondence that motivates the de Sitter branch of the thermodynamics."}],"review_version":1}