{"id":"faf87527-1014-424e-b6b5-d6e423902d07","arxiv_id":"1908.04426","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Benchmark neutron-matter calculations find that unconstrained auxiliary-field diffusion Monte Carlo agrees with Brueckner-Bethe-Goldstone theory, while constrained AFDMC overestimates the energy when spin-orbit forces are present.","lead":"This paper compares three independent computer methods for calculating how much energy packed neutrons hold at different densities, using several realistic models of the neutron-neutron force. It shows that a widely used constrained quantum Monte Carlo trick overestimates the energy when spin-orbit forces are present, and that relaxing the constraint brings the result into agreement with another method.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"AFDMC-UC agreement with BBG rests on extrapolation and 14-neutron additivity that are validated only for AV8' at ρ0; a direct high-density unconstrained run is needed.","rationale":"The paper is careful: the plus/minus importance sampling, the covariance-aware fitting procedure, and the 38-neutron AV8' check are genuine internal controls, and the AV8' unconstrained result is compared to GFMC. Those controls, however, do not cover the regime where the claimed correction is largest, namely the NV2 potentials at 2ρ0. The central comparison with BBG is only as strong as the two extrapolations identified by the reader. I do not think this warrants rejection; the trend is plausible and partially validated, but it does require either a direct high-density unconstrained calculation or release of the data and code needed to verify the additivity and extrapolation. The reader's CONDITIONAL verdict therefore remains appropriate.","tokens_in":24568,"tokens_out":6376,"duration_ms":69407,"concrete_test":"Run a direct 66-neutron unconstrained propagation for NV2-Ib at ρ = 2ρ0 extending τ to at least 0.008 MeV^-1, fit EUC(τ) with a single exponential and also with a two-exponential form, and compare the resulting E0 with the 14-neutron-based additive estimate used in Fig. 7. If the direct value differs from the additive estimate by more than the quoted statistical error, the AFDMC-UC equation of state at high density is not validated.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim (Secs. IV and V) that releasing the constrained-path bias brings AFDMC into agreement with BBG depends on two steps in Sec. III C: (i) a single-exponential fit to EUC(τ) over τ ≤ 0.004 MeV^-1 after Eq. (43), and (ii) adding the 14-neutron EUC(τ) - EUC(τ0) difference to 66-neutron AFDMC-CP energies (Sec. IV). The 38-neutron check is reported only for AV8', apparently at ρ = ρ0, and it tests the box-size independence of the difference, not the convergence of the exponential fit. Since the claimed bias grows strongly with density (up to ~8 MeV for NV2-Ib at 2ρ0), both approximations could fail exactly where the agreement with BBG is most needed. Moreover, the unconstrained propagation starts from constrained configurations in which negative-overlap components were discarded; a short single-exponential fit can miss a slow component, and the authors' own note that backflow correlations lower constrained energies by more than 1 MeV suggests residual trial-wave-function sensitivity. Thus the 3 MeV agreement is not yet established outside the validated AV8' ρ0 case.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports benchmark calculations of the energy per particle of pure neutron matter as a function of density, using three independent many-body methods: Brueckner–Bethe–Goldstone (BBG), Fermi hypernetted chain/single-operator chain (FHNC/SOC), and auxiliary-field diffusion Monte Carlo (AFDMC). The interactions considered are AV6', AV8', AV18, and four Norfolk NV2 chiral potentials. Two technical improvements are central: the FHNC/SOC calculation includes additional elementary diagrams, and the AFDMC calculation implements unconstrained imaginary-time propagation after a constrained-path stage, together with a new importance-sampling procedure. The main result is that for potentials with spin-orbit terms, constrained AFDMC significantly overestimates the neutron-matter energy, while unconstrained AFDMC lowers the energy and brings it into close agreement with BBG up to twice saturation density, with discrepancies of order 3 MeV per particle or less. The paper also finds that potentials fit to higher-energy NN scattering data (AV18, NV2-II) give a smaller spread of the equation of state than potentials fit only up to 125 MeV laboratory energy.","tokens_in":24770,"tokens_out":5831,"duration_ms":61753,"significance":"If the central result holds, it is significant for the nuclear many-body field: it identifies a systematic upward bias in previous constrained AFDMC neutron-matter equations of state at high density, and it demonstrates that releasing the constrained-path bias brings two very different methods, AFDMC and BBG, into agreement for the NN-only Hamiltonian. The benchmark is also useful because it covers two families of realistic potentials and provides a consistent comparison of methods. The paper has notable strengths: the AFDMC energies carry carefully estimated statistical errors from a covariance-matrix analysis of the imaginary-time data; the unconstrained-propagation correction is cross-checked with 14- and 38-neutron boxes for AV8'; the FHNC/SOC treatment is improved beyond previous work; and the phase-shift comparisons connect the density range of the equation of state to the laboratory-energy range of the scattering data. The potentials are external inputs fit to scattering data, so the benchmark comparison is not circular with respect to the reported neutron-matter energies.","major_comments":[{"comment":"The unconstrained AFDMC energies are constructed by adding the 14-neutron difference EUC(τ)−EUC(τ0) to the 66-neutron AFDMC-CP values. The box-size validation of this difference is reported only for AV8', and the text does not state that it was performed at densities other than ρ0 (the figure context suggests ρ0). The correction is then applied at all densities up to 2ρ0 and for all spin-orbit potentials, where it is as large as about 8 MeV per particle for NV2-Ib. Because the claimed agreement between AFDMC-UC and BBG depends directly on this additivity and on the single-exponential extrapolation, the manuscript should provide a direct unconstrained run at high density for at least one spin-orbit potential (for example AV8' or NV2-IIb with 38 or 66 neutrons) to show that the correction is not density-dependent in a way that invalidates the 3 MeV agreement.","section":"Sec. III C, Eq. (43) and Fig. 4"},{"comment":"The asymptotic energy E0 is obtained from a single-exponential fit to EUC(τ) over τ ≤ 0.004 MeV^-1. The unconstrained propagation starts from configurations produced by a constrained propagation, so components discarded by the constraint are not resampled at τ0; the authors also note that backflow correlations lower the constrained energy by more than 1 MeV per particle. These facts leave a residual trial-wave-function dependence that the single-exponential fit does not quantify. I request a stability check of E0 under a two-exponential fit and, if feasible, a test with backflow-correlated trial wave functions at ρ0 and at one high density, so that the central AFDMC-UC/BBG comparison rests on a more controlled extrapolation.","section":"Sec. III C, Eq. (43) and Fig. 4"}],"minor_comments":[{"comment":"The caption reads \"Same as Fig. 4\" but should refer to Fig. 3.","section":"Fig. 4 caption"},{"comment":"The sentence giving the AV18 constrained-minus-unconstrained difference states \"∼ 3 MeV at ρ = ρ0 and ∼ 7 MeV at ρ = ρ0\"; the second density should be 2ρ0.","section":"Sec. V"},{"comment":"The text contains the duplicated phrase \"Since the the expectation values are substantially correlated in τ\"; the second \"the\" should be removed.","section":"Sec. III C"},{"comment":"The symbol α is used both for the spin/isospin potential quencher in Eq. (25) and for the parameter appearing in the positive-definite guiding function in Eq. (43); renaming one of them would avoid confusion.","section":"Eq. (25) and Eq. (43)"},{"comment":"The conclusion that potentials fit to higher-energy scattering data reduce the spread of the equation of state is based on four of the six NV2 variants; including the NV2-Ic and NV2-IIc cutoffs would make the regulator dependence of this conclusion quantitative rather than illustrative.","section":"Sec. V"},{"comment":"The BBG results are presented without uncertainty estimates; given that the AFDMC-UC/BBG agreement is a central benchmark, a remark on the expected size of the hole-line truncation error, for example from three-hole-line contributions, would help the reader interpret the \"within 3 MeV\" claim.","section":"Sec. IV"}],"recommendation":"major_revision","confidential_remarks":"The stress-test concern raised in the reader's note is real: the AFDMC-UC/BBG agreement rests on an additivity and extrapolation procedure that is directly validated only for AV8' at normal density, while the correction is largest at high density. The paper is otherwise careful and the central finding is likely correct, but the missing high-density validation is load-bearing for the main claim and should be addressed in revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The short version: this is a careful benchmark that delivers a new and consequential result—constrained AFDMC neutron-matter energies carry a spin-orbit-dependent bias that grows with density, and releasing the constraint moves AFDMC close to BBG. I think the authors have earned that conclusion, with one caveat about how much weight the extrapolation can carry.\n\nWhat's genuinely new: the plus/minus importance sampling applied to pure neutron matter, unconstrained propagation with a proper covariance-aware exponential fit, and the extension of FHNC/SOC with central elementary diagrams beyond FHNC/4. The paper also does a useful job connecting the many-body spread to how well each potential describes NN phase shifts at the relevant lab energies. The AV18/NV2 comparison is informative, and the point that potentials fit to higher-energy scattering data give a tighter EoS band is well supported.\n\nThe central comparison is internally consistent. The unconstrained runs visibly lower the energy for AV8' (about 3 MeV at rho0), the 38-neutron check for AV8' supports box-size independence of the difference, and the AFDMC-UC numbers land within a couple MeV of BBG across most of the range. That is a satisfying convergence of methods.\n\nThe soft spots are concentrated in the machinery that produces AFDMC-UC. The asymptotic energy comes from a single-exponential fit to EUC(tau) over a short propagation window, and the correction is computed with 14 neutrons and added to 66-neutron constrained results. That additivity and the fit quality are validated in detail only for AV8' at rho0. The density where the bias matters most—around 2 rho0, where the claimed shift is up to ~8 MeV for NV2-Ib—is exactly where the validation is thinnest. The authors are honest that backflow correlations change constrained results by more than 1 MeV, which is a reminder that trial-wave-function sensitivity has not been fully eliminated. BBG results also come without uncertainty estimates, so the 'agreement' has a one-sided error bar.\n\nNone of this is fatal. The paper is transparent about its approximations, the statistics on AFDMC-UC are properly handled, and the qualitative conclusion—constrained AFDMC is biased upward for potentials with spin-orbit terms—is robust. The quantitative 1–2.5 MeV agreement with BBG is best read as provisional until a direct high-density unconstrained calculation with a larger box is done.\n\nWho gets value: anyone using or comparing microscopic neutron-matter EoS, especially in the neutron-star context, and people developing QMC methods. It deserves a serious peer review, and I would want the high-density extrapolation checked in review.","headline":"The paper makes a real, mostly convincing case that constrained-path AFDMC systematically overestimates neutron-matter energies once spin-orbit interactions are in the potential, and that unconstrained propagation fixes most of the discrepancy; treat the high-density end as provisional.","tokens_in":25365,"tokens_out":3627,"would_cite":true,"duration_ms":33247,"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":"In pure neutron matter, constrained-path Monte Carlo overestimates the energy when the nucleon-nucleon force has spin-orbit terms; releasing the constraint removes the bias and matches Brueckner theory.","keywords":["pure neutron matter","equation of state","auxiliary-field diffusion Monte Carlo","Brueckner-Bethe-Goldstone","Fermi hypernetted chain","spin-orbit interaction","constrained-path approximation","nucleon-nucleon potentials"],"falsifier":"Run the unconstrained stage directly with 66 neutrons at $\\rho_0$ and $2\\rho_0$ for the AV8′, AV18, and one NV2 potential, and compare with the values obtained by adding the 14-neutron unconstrained correction to the 66-neutron constrained results; a difference larger than the quoted error bars would break the central agreement with Brueckner theory. A complementary check is to compute the same periodic system with a method that treats spin-isospin exactly, such as Green's function Monte Carlo, at the densities where the correction is largest.","tokens_in":24323,"feed_emoji":"⚛️","tokens_out":15958,"duration_ms":146653,"temperature":0.7,"pith_summary":"The paper benchmarks three independent many-body methods—Brueckner–Bethe–Goldstone, Fermi hypernetted chain/single-operator chain, and auxiliary-field diffusion Monte Carlo—for the energy per particle of pure neutron matter, using the AV6′, AV8′, and AV18 potentials and four NV2 chiral potentials. Its central finding is that the constrained propagation normally used in the Monte Carlo method to control the fermion sign problem is not neutral: for every potential with spin-orbit terms, the constrained energy per particle sits above the true result, and the gap grows with density. When the constraint is released and the imaginary-time evolution is extrapolated to its asymptotic energy, the Monte Carlo results drop substantially and come into close agreement with the Brueckner method, within about 1 MeV for AV18 and within about 2.5 MeV for the NV2 family up to twice nuclear saturation density. The paper also reports that the hypernetted-chain method lies below the other two above saturation density, an effect attributed to the truncated treatment of spin-orbit correlations. If the central claim is right, earlier constrained Monte Carlo equations of state for neutron matter are systematically too stiff at high density, and potentials fitted to higher-energy scattering data give a tighter band of predictions.","feed_headline":"Unconstrained Monte Carlo: neutron-matter energy drops by up to 8 MeV","feed_subtitle":"Releasing the constrained propagation removes the high-density bias in neutron-matter equations of state.","key_machinery":"The central object is a two-stage imaginary-time evolution in auxiliary-field diffusion Monte Carlo. In the first stage the walkers are propagated with the constrained-path approximation, which suppresses the fermion sign problem but is not variational. In the second stage the constraint is released: the guiding function is switched to $\\Psi_G(X)=\\sqrt{\\mathrm{Re}\\{\\Psi_T(X)\\}^2+\\alpha\\,\\mathrm{Im}\\{\\Psi_T(X)\\}^2}$ with $\\alpha=0.5$, and the energy $E_{\\rm UC}(\\tau)$ is tracked until it can be extrapolated to its asymptotic value by a single-exponential fit that accounts for correlations through the covariance matrix. The difference $E_{\\rm UC}(\\tau)-E_{\\rm UC}(\\tau_0)$ computed with 14 neutrons in a periodic box is added to the 66-neutron constrained results to estimate the thermodynamic-limit unconstrained energy; the additivity is validated for AV8′ by comparing 14- and 38-neutron boxes. This release-and-extrapolate mechanism carries the paper's central claim, because it is what turns the biased constrained energies into the energies that agree with Brueckner theory.","core_discovery":"The central claim is that the constrained-path approximation in auxiliary-field diffusion Monte Carlo introduces a density-dependent upward bias in the neutron-matter equation of state whenever the two-nucleon interaction contains spin-orbit terms, and that removing this bias reconciles the method with Brueckner theory. Concretely, at $\\rho_0$ the AV8′ energy per particle falls from $15.55(1)$ MeV under the constrained propagation to $12.5(3)$ MeV after the constraint is released; for AV18 the release lowers the energy by about $2.2$ MeV at $\\rho_0$ and $5.3$ MeV at $2\\rho_0$, and for NV2-Ib the shift reaches about $8$ MeV at $2\\rho_0$. The paper presents this as a cross-method benchmark: after the correction, AFDMC and BBG agree within $1$ MeV for AV18 and within about $2.5$ MeV for the NV2 potentials up to $2\\rho_0$, while FHNC/SOC falls below the other two at supra-saturation densities because its spin-orbit correlations are truncated at the three-body cluster level. The authors frame the result as a systematic improvement to the Monte Carlo method rather than a final equation of state, since three-nucleon forces are deliberately excluded.","pith_inferences":["A likely consequence: earlier constrained AFDMC equations of state for pure neutron matter obtained with local N2LO chiral Hamiltonians carry the same upward spin-orbit bias; running the unconstrained stage for those interactions would test this directly.","A diagnostic extension: if the bias is driven by spin-orbit strength, its size should correlate with the splitting of the $^3P_J$ phase shifts across the six potentials used here, which could be checked at fixed density.","A practical take-home: for neutron-star modelling, two-body-only constrained equations of state should be treated as upper bounds above saturation density, with the unconstrained correction and three-nucleon forces both lowering the energy."],"forward_implications":["For potentials with spin-orbit terms, constrained AFDMC overestimates the energy per particle, and the bias grows with density: about $2.2$ MeV at $\\rho_0$ and $5.3$ MeV at $2\\rho_0$ for AV18, and up to about $8$ MeV at $2\\rho_0$ for NV2-Ib.","After the constraint is released, AFDMC and BBG agree within about $1$ MeV per particle for AV18 and within about $2.5$ MeV for the NV2 potentials up to twice nuclear saturation density.","The FHNC/SOC energies fall below both BBG and AFDMC-UC above saturation density, and the paper attributes this to the three-body cluster truncation of spin-orbit correlations.","Potentials fitted to higher-energy scattering data (AV18, NV2-IIa, NV2-IIb) keep the spread of equations of state within about $4$ MeV per particle up to $2\\rho_0$, while including the lower-energy-fitted NV2-I models widens the spread to about $9$ MeV.","For the spin-orbit-free AV6′ potential, all three methods agree within about $5$ MeV per particle up to $2\\rho_0$, and constrained and unconstrained AFDMC nearly coincide."],"supporting_citations":[{"why":"Defines the AV18 potential, one of the two Hamiltonian families whose equations of state are benchmarked.","marker":"[56]"},{"why":"Defines the simplified AV8′ and AV6′ potentials; the AV8′ case demonstrates the spin-orbit constraint bias.","marker":"[57]"},{"why":"Establishes the local chiral two-nucleon framework in which the NV2 potentials are constructed.","marker":"[58]"},{"why":"Provides the complete NV2 interactions with N3LO contact terms used in the benchmark.","marker":"[59]"},{"why":"The earlier cross-method benchmark whose discrepancies motivate this work, and the source of the spin-orbit sensitivity hypothesis.","marker":"[60]"},{"why":"Supplies the plus-and-minus importance sampling adopted in the AFDMC calculations.","marker":"[61]"},{"why":"Provides the complex-wave-function constrained-path algorithm on which the constrained propagation is based.","marker":"[105]"},{"why":"Introduces the positive-definite guiding function used for the unconstrained propagation stage.","marker":"[108]"},{"why":"Earlier Green's function Monte Carlo neutron-matter results whose consistency supports the unconstrained correction for AV8′.","marker":"[115]"},{"why":"Previous constrained AFDMC neutron-matter equation of state used as the baseline that the unconstrained results correct.","marker":"[116]"}],"fun_headline_variants":["Neutron matter energy drops when Monte Carlo constraint is lifted","AFDMC bias fixed: neutron-matter energy falls up to 8 MeV","Releasing constraint reconciles neutron-matter methods","Spin-orbit bias removed in neutron-matter Monte Carlo","Benchmark: unconstrained AFDMC tightens neutron-matter energies"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the energy lowering measured during unconstrained evolution of 14 neutrons in a small periodic box is exactly the correction that applies to the 66-neutron constrained energies at every density and for every potential, and that the extrapolation used to read off that lowering is trustworthy.","fun_headline_variants_meta":{"raw":{"variants":["Neutron matter energy drops when Monte Carlo constraint is lifted","AFDMC bias fixed: neutron-matter energy falls up to 8 MeV","Releasing constraint reconciles neutron-matter methods","Spin-orbit bias removed in neutron-matter Monte Carlo","Benchmark: unconstrained AFDMC tightens neutron-matter energies"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00013,"raw_usage":{"total_tokens":1134,"prompt_tokens":961,"completion_tokens":173,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":577,"completion_tokens_details":{"reasoning_tokens":85}},"tokens_in":577,"tokens_out":173,"duration_ms":3019,"temperature":1.0,"reasoning_tokens":85,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:43:22.344090+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the unconstrained stage directly with 66 neutrons at $\\rho_0$ and $2\\rho_0$ for the AV8′, AV18, and one NV2 potential, and compare with the values obtained by adding the 14-neutron unconstrained correction to the 66-neutron constrained results; a difference larger than the quoted error bars would break the central agreement with Brueckner theory. A complementary check is to compute the same periodic system with a method that treats spin-isospin exactly, such as Green's function Monte Carlo, at the densities where the correction is largest.","supporting_citations":[{"cited_title":"Comparative study of neutron and nuclear matter with simplified Argonne nucleon-nucleon potentials","cited_arxiv_id":"1207.6314","evidence_quote":"Supplies the plus-and-minus importance sampling adopted in the AFDMC calculations."},{"cited_title":"Optimization of quantum Monte Carlo wave functions by energy minimization","cited_arxiv_id":"physics/0701039","evidence_quote":"Provides the complex-wave-function constrained-path algorithm on which the constrained propagation is based."},{"cited_title":"Spin-orbit induced backflow in neutron matter with auxiliary field diffusion Monte Carlo","cited_arxiv_id":"nucl-th/0304042","evidence_quote":"Earlier Green's function Monte Carlo neutron-matter results whose consistency supports the unconstrained correction for AV8′."},{"cited_title":"Quantum Monte Carlo Calculations of Neutron Matter","cited_arxiv_id":"nucl-th/0302041","evidence_quote":"Previous constrained AFDMC neutron-matter equation of state used as the baseline that the unconstrained results correct."}],"review_version":1}