{"id":"487e77e3-8273-4af4-87c3-29871883eb0d","arxiv_id":"2606.16385","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Universal machine-learning interatomic potentials fail standard melt-quench amorphous-structure generation, expanding IrO2 to gas-like densities; a fixed-volume quench plus low-temperature NPT equilibration restores AIMD-matching densities across 30 materials.","lead":"This paper tests eight popular 'universal' machine-learning atomic models on the standard melt-quench recipe for making amorphous materials and finds that all of them expand the structure into a low-density gas. It then shows that a simple change—quenching at fixed volume and relaxing pressure only at low temperature—restores correct densities for 30 materials, matching ab initio results.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Revised-protocol density recovery may be volume-heredity artifact: final NPT is only 10 ps and initial cell volume is set near crystal density, so the reported MAE improvement could reflect box history, not corrected E-V/pressure physics.","rationale":"The reader's weakest assumption concerns mixed AIMD/experimental/crystal reference densities in the 30-material benchmark. That is a real issue, but the more fundamental threat to the paper's central claim is that the revised protocol's success may be an artifact of volume heredity. Because the NVT stage fixes the cell at a crystal-density-based volume, and because the final NPT stage is only 10 ps, the reported densities may not represent equilibrium under the uMLIP Hamiltonian. For potentials with monotonically decreasing E-V curves, the only reason they do not expand during the revised protocol is that the high-temperature volume-expanding regime is bypassed and the low-temperature equilibration is too short to reveal the missing energy minimum. This does not overturn the empirical NPT-failure observation for IrO2, nor the useful diagnostic that E-V and pressure errors matter. However, it means the paper's proposed 'practical solution' is not yet validated as a physical fix. The appropriate verdict remains CONDITIONAL: the manuscript should be accepted only if the revised-protocol density is shown to be converged and independent of initial volume. The reference-comparability issue raised by the reader is partially complementary: both concerns converge on the 30-material benchmark, but the volume-heredity issue is more direct and more damaging to the central 'restores density' claim.","tokens_in":18751,"tokens_out":6714,"duration_ms":78962,"concrete_test":"Run the revised protocol for IrO2 with GRACE and UPET starting from three initial volumes: 0.8, 1.0, and 1.2 times the rutile cell volume; for representative 30-set materials (e.g., SiO2, In2O3), use Packmol packings at 0.9, 1.05, and 1.2 times crystal density. Extend the final 300 K NPT equilibration from 10 ps to at least 200 ps, and also initialize a 300 K NPT run from a deliberately expanded 4 g/cm3 IrO2 configuration. If final densities converge to ~10.0 g/cm3 for IrO2, independent of starting volume and stable for 1 ns, the protocol is physically sound. If densities track the initial volume, keep drifting, or remain gas-like when started from low density, the reported recovery is a box-history artifact.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central remedy—NVT-quench followed by short NPT equilibration—may not actually restore equilibrium densities because the final NPT stage is too short and the starting volume is predetermined by the crystal. IrO2 simulations begin from the rutile crystal cell; the 30-material benchmarks begin from Packmol packings deliberately set to 1.05 times the Materials Project crystal density. The NVT melt/quench holds that cell volume fixed, so the density entering the final NPT stage is set by the crystal-based box, not by the uMLIP. The final NPT equilibration is only 10 ps with an ASE Berendsen-style barostat, and no convergence or box-size/history dependence is reported. This is load-bearing because for GRACE, ORB, SevenNet, MACE, MatterSim, and UMA, the E-V curves in Fig. 3c–h are monotonically decreasing with density—i.e., there is no volume-energy minimum. A true NPT simulation at 300 K should drift to lower density if given enough time. The observed ~10 g/cm3 may therefore be kinetic trapping at the initial crystal-like volume, and the headline reduction from 2.46 to 0.35 g/cm3 MAE may reflect starting every box near the reference density rather than a physical cure. The NPT-failure diagnosis for IrO2 stands, but the claim that the revised protocol recovers AIMD-consistent densities is not yet established.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper evaluates eight universal machine-learning interatomic potentials (GRACE, ORB, SevenNet, MACE, MatterSim, UMA, UPET, NequIP) for melt–quench generation of amorphous IrO2. It reports that all eight fail under the conventional NPT quench protocol, yielding densities of 1–4 g/cm3 versus an AIMD (PBE) reference of 10.04 g/cm3. Using DFT benchmarks on trajectory snapshots, the authors argue that the failure is caused by incorrect energy–volume responses and/or virial-pressure overestimation, rather than by poor pointwise energy/force accuracy. They propose two remedies: pressure-targeted fine-tuning of UPET, and a revised NVT-quench/NPT-equilibration protocol. They claim the revised protocol restores IrO2 densities for all eight models and reduces the density MAE from 2.46 to 0.35 g/cm3 on a 30-material benchmark (AIMD-referenced subset, GRACE runs).","tokens_in":1690,"tokens_out":1494,"duration_ms":116703,"significance":"If the central claims hold, this is a useful and timely contribution to a practical problem: universal MLIPs are increasingly used for amorphous structure generation, and standard energy/force benchmarks are shown to be insufficient for NPT melt–quench simulations. The diagnosis that the sign of the E–V error and the fidelity of virial pressure are controlling factors is more informative than aggregate energy/force MAEs alone. Strengths include repeated independent IrO2 runs, DFT benchmarking of trajectory snapshots, an eight-model comparison on IrO2, and an out-of-sample 30-material test of the revised protocol. However, the 30-material test uses only one uMLIP (GRACE), mixes reference types of unequal physical meaning, and is subject to a volume-heredity concern that requires additional convergence evidence before the protocol can be accepted as a general physical cure.","major_comments":[{"comment":"The central claim that the revised NVT-quench/NPT-equilibration protocol recovers AIMD-consistent densities is not yet established because the initial cell volume is set externally and the final NPT stage is very short. For IrO2, the simulations start from the rutile crystal cell; for the 30-material set, initial packings are built at 1.05× the Materials Project crystal density. Since the quench is performed in NVT, the density entering the 10 ps NPT equilibration is predetermined to be near the crystal density. For GRACE, ORB, SevenNet, MACE, MatterSim, and UMA, the E–V curves in Fig. 3c–h are monotonically decreasing over the tested range, so no stable volume minimum exists at 300 K; a longer NPT run would be expected to continue expanding. The paper reports no density-versus-time curves for the final NPT stage, no longer-NPT tests, and no dependence on the starting cell volume. The re","section":"§2.4 and Methods §4.1; Fig. 6c–e"},{"comment":"The 30-material benchmark mixes reference types and uses only GRACE. The headline AIMD-referenced MAE includes the LDA-based WO3 entry, and 11 of the 30 entries use experimental or crystalline-phase reference densities. For amorphous structure generation, a crystalline reference density is not a commensurate ground truth: amorphous densities are generally lower than crystal densities, and the revised protocol deliberately starts from crystal-density packings, so a model that stays near the starting volume will appear to succeed against crystal references even if its amorphous density is wrong. In addition, all 30-material runs are performed with only one uMLIP (GRACE), so the abstract/conclusion statements that the failure and remedy are general to uMLIPs are broader than the evidence. Please report per-reference-source metrics, remove or clearly separate non-PBE-AIMD entries from the he","section":"Table S2 and Fig. 6e–f"},{"comment":"The pressure-targeted fine-tuning demonstration is an in-domain retraining check rather than an independent validation. UPET-FT is trained on AIMD IrO2 configurations and then evaluated on the same material, so it shows that a uMLIP can be corrected when abundant target-domain AIMD data are available, but it does not establish fine-tuning as a general remedy. If this is retained as one of the two central practical solutions, please either restrict the claim to a proof-of-concept or add a second uMLIP and a held-out material to demonstrate transferability.","section":"§2.3, Figs. 4–5"}],"minor_comments":[{"comment":"The loss expression contains '10.0&&' where the stress-loss coefficient should be defined. Please replace with an explicit term such as 10.0 L_S.","section":"Methods, Eq. (4)"},{"comment":"The density values for the revised protocol are reported to two decimals and agree closely with AIMD, but no error bars are visible for the individual uMLIP bars in the figure as printed. Please clarify the statistics for these revised-protocol runs.","section":"Fig. 6c"},{"comment":"The pressure control uses the ASE Berendsen-style barostat. Since the paper’s central concern is NPT volume dynamics, please state whether the conclusions are expected to hold with a more rigorous barostat (e.g., Parrinello–Rahman) and whether any thermostat/barostat coupling parameters were varied.","section":"Methods §4.1"},{"comment":"The MTP used as a domain-specific reference is cited as 'In preparation' (Ref. 25). For reproducibility, the MTP training data, active-learning details, and final potential should be made available in a public repository.","section":"Data availability"}],"recommendation":"major_revision","confidential_remarks":"The central IrO2 failure diagnosis is convincing and likely of broad interest. The main uncertainty is whether the revised protocol’s reported density recovery is a genuine equilibrium property of the uMLIPs or a short-time volume-heredity artifact; this is fixable with additional convergence simulations and a cleaner reference set. I do not see grounds for rejection, but the load-bearing generalization claim needs strengthening before publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Worth reading, but the headline fix is less proven than the authors claim. The genuinely new and valuable part is the systematic demonstration that eight uMLIPs all fail under the conventional NPT melt-quench protocol for amorphous IrO2, plus the physical diagnosis: accurate energies and forces alone do not guarantee stable NPT dynamics, and the sign of the energy-volume error and the fidelity of virial pressure matter. The per-snapshot benchmarking against DFT and the E-V curves are well done, and the IrO2 failure is convincing: repeated runs, DFT/AIMD checks, and a system-specific MTP that behaves correctly as contrast. The pressure-targeted fine-tuning of UPET restoring density and RDF for IrO2 is also a useful proof of concept, though it is an in-domain retraining check on AIMD configurations rather than an independent prediction.\n\nThe soft spot is the 30-material generalization and the revised NVT-quench/NPT-equilibration protocol. The final NPT stage is only 10 ps, and the initial volume is set near the crystal density—IrO2 starts from the rutile cell, and the Packmol packings use 1.05 times the Materials Project crystal density. For several of the uMLIPs, the E-V curve is monotonic, with no volume-energy minimum. In a genuine NPT equilibrium those models should keep expanding; 10 ps with a Berendsen barostat is not long enough to show they don't. The reported MAE drop from 2.46 to 0.35 g/cm3 may therefore be mostly box history—kinetic trapping at the starting volume—rather than a physical cure. The paper does not report convergence with respect to NPT duration, nor volume-history dependence. The heterogeneous reference set in Table S2 (AIMD, experiment, crystal, one LDA value) adds a further layer of uncertainty. No code or inputs are shipped, and the successful system-specific MTP is an in-preparation manuscript, so several numerical claims cannot currently be checked.\n\nThese issues do not undermine the failure diagnosis. Anyone doing melt-quench with uMLIPs should read this and adopt the suggested validation criteria: check density, E-V response, and pressure, not just energy/force errors. But the claim that the revised protocol is a general remedy is not yet established.\n\nRecommendation: send to peer review with a serious referee, and require longer NPT equilibration tests, volume-history checks, and either code/data release or detailed convergence evidence before publication.","headline":"Convincing failure diagnosis for uMLIP melt-quench, but the claimed general cure is under-supported—the short NPT stage likely inherits the starting crystal-like volume.","tokens_in":19591,"tokens_out":2949,"would_cite":true,"duration_ms":35947,"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":"Universal machine-learning interatomic potentials fail at melt-quench amorphous structure generation because their energy-volume and pressure errors cause unphysical expansion; revised NVT quench and pressure-targeted fine-tuning restore re","keywords":["universal machine-learning interatomic potentials","melt-quench molecular dynamics","amorphous IrO2","energy-volume response","virial pressure","NVT-quench protocol","pressure-targeted fine-tuning","amorphous structure generation"],"falsifier":"Rerun the conventional and revised melt-quench protocols with any of the eight potentials on a material with an accurately known experimental amorphous density outside the 30-material set (e.g., fused silica or amorphous selenium). If the conventional protocol produces a near-reference density, or the revised protocol misses it by more than the reported error scale, the generality claim is wrong. Alternatively, compute the DFT energy along one of the uMLIP's expansion paths: if DFT shows an energy minimum near the reference density where the potential predicts a monotonic decrease, the propose","tokens_in":18574,"feed_emoji":"⚛️","tokens_out":6973,"duration_ms":63920,"temperature":0.7,"pith_summary":"The paper tries to establish that the conventional NPT melt-quench protocol, when driven by eight leading universal machine-learning interatomic potentials, produces unphysically expanded amorphous structures, not because energies or forces are inaccurate but because the models misjudge how energy changes with volume and how virial pressure responds. Using amorphous IrO2 as a diagnostic, it shows all eight models predict densities of 1-4 g/cm3 against a DFT reference of 10.04 g/cm3. The paper then demonstrates two fixes: fine-tuning with a stress-weighted loss, and, more generally, a revised protocol that keeps the cell volume fixed during the high-temperature quench and only applies NPT at low-temperature equilibration. Across 30 chemically diverse materials, the revised protocol reduces the mean absolute density error from 2.46 to 0.35 g/cm3 against AIMD references. The significance is that validation of universal potentials must include energy-volume curves and pressure, not just energy and force benchmarks.","feed_headline":"Revised quench protocol fixes amorphous densities across 30 materials","feed_subtitle":"Melt-quench with universal ML potentials expands oxides to gas-like states; revised NVT quench restores real densities.","key_machinery":"The load-bearing objects are the energy-volume (E-V) curve and the virial pressure computed from the stress tensor. The paper's central diagnostic is the sign of the E-V error under volume expansion, not its magnitude, and the central fix is a simulation protocol in which the high-temperature quench runs in the NVT ensemble (fixed cell volume) and only the low-temperature equilibration runs in NPT, so incorrect high-temperature E-V or pressure responses cannot drive volume evolution. A second mechanism is pressure-targeted fine-tuning, which reweights the loss toward the stress term to repair pressure predictions without losing energy/force accuracy.","core_discovery":"The central claim is that accurate energies and forces are necessary but not sufficient for stable NPT melt-quench dynamics; what matters is the sign of the energy-volume error and the fidelity of the virial pressure. The paper shows that a model can be more accurate than a system-specific reference on every instantaneous quantity and still drive catastrophic expansion, while a less accurate model with the correct E-V sign stays stable. It also shows that even models whose energy-volume curves match DFT can still fail because they systematically overestimate pressure. The two remedies — pressure-targeted fine-tuning and an NVT-quench/NPT-equilibration protocol — both recover amorphous IrO2 d","pith_inferences":["If the sign-of-E-V-error mechanism is general, uMLIP training sets should deliberately include expanded, low-density amorphous configurations so the models learn the correct energetic penalty for volume growth.","The revised protocol fixes density by suppressing volume degrees of freedom at high temperature; properties that depend on genuine high-temperature volume response, such as thermal expansion or pressure-induced transformations, would still require pressure-accurate models.","The 30-material benchmark mixes AIMD, experimental, and crystalline reference densities; the qualitative failure is robust, but the quantitative 0.35 g/cm3 MAE is only defined on the 19 AIMD-referenced materials, and extending it to all 30 assumes those references are comparable.","A direct extension would be to test whether the same NVT-quench fix applies to other non-equilibrium NPT processes, such as rapid cooling of metallic glasses or nucleation from the melt, where the same pressure-error mechanism could distort the outcome."],"forward_implications":["Benchmarks for universal machine-learning interatomic potentials should include energy-volume curves and pressure parities, not just energy and force errors, because the conventional metrics miss the failure mode.","Published amorphous structures generated with universal potentials via conventional NPT melt-quench should be re-examined; the expanded densities are an artifact of the protocol, not the material.","The revised NVT-quench/NPT-equilibration protocol is a drop-in fix that requires no additional training data and should become the default for uMLIP-driven amorphous structure generation.","When fine-tuning is feasible, stress-weighted fine-tuning can correct volume behavior for a specific material while preserving energy and force accuracy.","The failure generalizes across oxides, nitrides, sulfides, halides, and ternary oxides, so any uMLIP used for non-equilibrium or volume-varying simulations should first pass an E-V/pressure validation."],"fun_headline_variants":["Melt-quench ML failures fixed by revised protocol across 30 materials","Melt-quench blowup from ML potentials fixed by pressure-targeted tuning","Why accurate ML forces aren't enough: melt-quench needs pressure fix","Revised NVT quench saves melt-quench from ML potentials' volume blowup"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The broad claim of general applicability rests on treating the 30 reference densities — some from computer simulation, some from experiment, some from the corresponding crystals — as fair targets for the potential; if those references are not directly comparable with what the simulation is trying to reproduce, the reported improvement is partly an artifact of the reference list.","fun_headline_variants_meta":{"raw":{"variants":["Melt-quench ML failures fixed by revised protocol across 30 materials","Melt-quench blowup from ML potentials fixed by pressure-targeted tuning","Why accurate ML forces aren't enough: melt-quench needs pressure fix","Revised NVT quench saves melt-quench from ML potentials' volume blowup"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000648,"raw_usage":{"total_tokens":2833,"prompt_tokens":784,"completion_tokens":2049,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":528,"completion_tokens_details":{"reasoning_tokens":1976}},"tokens_in":528,"tokens_out":2049,"duration_ms":16364,"temperature":1.0,"reasoning_tokens":1976,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-02T11:11:21.681571+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Rerun the conventional and revised melt-quench protocols with any of the eight potentials on a material with an accurately known experimental amorphous density outside the 30-material set (e.g., fused silica or amorphous selenium). If the conventional protocol produces a near-reference density, or the revised protocol misses it by more than the reported error scale, the generality claim is wrong. Alternatively, compute the DFT energy along one of the uMLIP's expansion paths: if DFT shows an energy minimum near the reference density where the potential predicts a monotonic decrease, the propose","supporting_citations":[],"review_version":1}