{"id":"e5df8632-20f0-4fe8-83d9-aee58e03d3c2","arxiv_id":"2506.18835","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Simulations with the OPLS/AA force field and DFT validation show that a linear PEI chain with ten amino groups forms stable complexes with up to four Hg2+ ions in water.","lead":"This paper uses atomic-scale computer simulations to show how a single chain of polyethyleneimine (PEI) with ten nitrogen atoms can hold onto up to four mercury ions at once. It matters for designing water-treatment materials that remove toxic mercury from wastewater.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The capacity claim rests on unvalidated Hg2+-amine cross-interactions; the MD Hg-N distance (~2.15 Å) disagrees with the paper's own DFT value (~2.5 Å), so the four-ion maximum should be treated as conditional until that interaction is re-benchmarked.","rationale":"I considered the pre-formation bias in Stage 1 of the protocol: starting from close-contact complexes tests stability, not spontaneous assembly. This is a real limitation, but the five-ion control (also pre-formed) dissociates, so the upper bound is not trivially explained by the bias. The force-field calibration issue is more load-bearing because it directly controls the coordination number and because the manuscript contains internal evidence (MD 2.15 Å vs DFT 2.5 Å) that the Hg-N interaction is not well represented. The DFT section is meaningful evidence that the selected geometries are local minima, but it cannot benchmark the classical Hg-N potential; starting geometries come from the MD model and no independent Hg-amine reference is provided. I do not think the paper should be rejected: the qualitative picture (PEI-10 binds multiple Hg2+, capacity somewhere near four) is plausible and the DFT-optimized structures are consistent with stable multi-ion complexes. But the numerical claim of exactly four should remain conditional until the Hg-N interaction is benchmarked and the ladder is re-run with a corrected model.","tokens_in":15016,"tokens_out":6145,"duration_ms":77047,"concrete_test":"Re-parameterize and re-run the capacity ladder. (1) Compute a reference Hg2+-amine potential using the same M06-2X/LanL2DZ/6-31+G(d,p) level for Hg2+ plus one and two methylamine ligands in implicit water, recording R(Hg-N) and binding energies. (2) Reproduce these with OPLS/AA + Babu-Lim Hg2+; if the MD minimum is below 2.4 Å or the binding energy differs by more than about 4 kcal/mol per bond, adjust the Hg epsilon/sigma (or add a Hg-N-specific cross term) until the amine benchmark matches. (3) Repeat the PEI-10 + n Hg2+ protocol for n=1-5 with the recalibrated parameters. If the recalibrated model still gives stable complexes for n=4 and dissociation for n=5, the central claim survives; otherwise the four-ion capacity is a force-field artifact.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The capacity claim ('up to four Hg2+ per PEI-10') is only as secure as the Hg2+-amine part of the force field. The Hg2+ parameters are taken from Babu-Lim, which were fitted to Hg2+ hydration, and all cross terms use the geometric combination rule; no comparison to known Hg2+-amine structures or binding energies is given. The paper itself contains a warning sign: MD gives a first Hg-N peak at ~2.15 Å (Section 4), while the DFT-optimized Hg-N distances in Table 2 are all about 2.48-2.57 Å. A 0.35 Å error in the first coordination shell is not minor: the MD Coulomb term between Hg2+ and the OPLS amine N (q=-0.78) is extremely strong at 2.15 Å, so the model likely overbinds Hg2+ to nitrogen. If the true Hg-N well is shallower or sits further out, the running coordination numbers n(r) in Figs. 4, 7, and 8 would change, and the maximum stable number of ions could be less than four. DFT does not repair this: the DFT optimizations start from MD-selected conformations and therefore test local stability of those topologies, not whether the force field produces the correct Hg-amine interaction. The central claim therefore rests on an unvalidated cross-interaction, with internal evidence that it is miscalibrated.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports atomistic molecular dynamics simulations of linear polyethyleneimine chains (PEI-4, PEI-5, PEI-10) complexed with Hg2+ ions in explicit SPC/E water, using OPLS/AA parameters for PEI and Babu-Lim parameters for Hg2+. Complexes are pre-formed in vacuum (Stage 1), solvated and equilibrated for 3 ns, then simulated for 4 ns in the production stage. The authors compute Hg-N, Hg-Ow, N-N, and Hg-Hg radial distribution functions and coordination numbers, and they supplement the MD with M06-2X/LanL2DZ DFT geometry optimizations, adsorption energies, and Gibbs free energies of complexation. The central claim is that one linear PEI-10 chain can coordinate up to four Hg2+ ions, with five ions being unstable.","tokens_in":15276,"tokens_out":5453,"duration_ms":61053,"significance":"If established, the predicted capacity of four Hg2+ ions per PEI-10 chain is a concrete, design-relevant quantity for PEI-based water-treatment adsorbents, and the paper usefully combines MD structural analysis with DFT checks. The study is also clearly written and parameter-free in the sense that no new force-field parameters are fitted. However, the central claim is not yet supported at the required level because (i) the Hg2+-amine cross-interaction is unvalidated and internally inconsistent with the paper's own DFT distances, (ii) all simulations start from pre-formed complexes, and (iii) all structural numbers come from single short trajectories without error estimates. These issues are fixable in principle, so the manuscript merits major revision rather than rejection.","major_comments":[{"comment":"The preparation protocol places the Hg2+ ion 'in close proximity' to the PEI molecule and runs Stage 1 in vacuum, so the production runs test only whether pre-formed Hg-N contacts remain bound, not whether complexes assemble from unbound components. Since the abstract's capacity claim ('capable of coordinating up to four Hg2+ ions') is about what a chain can do in solution, the manuscript should add simulations initialized from unbound or randomly placed ions (and from partially bound states) and show that the four-ion complex forms reproducibly. Without such a test, the 'up to four' statement could reflect kinetic trapping rather than thermodynamic capacity.","section":"Section 3 (Stage 1) and Section 4"},{"comment":"The MD Hg-N first peak is reported at approximately 2.15 Å (Figs. 4, 7, 8), while all DFT-optimized Hg-N distances in Table 2 are 2.48-2.57 Å. This roughly 0.35 Å discrepancy is internal evidence that the Hg2+-amine Lennard-Jones interaction, built from Babu-Lim hydration-fitted parameters (Table 1, ref. 36) via the geometric combination rule (Section 3), is miscalibrated for Hg-amine binding. Because the DFT optimizations start from MD-selected conformations, they do not repair this issue. The authors should benchmark the Hg2+-amine model against known Hg2+-amine complexes or high-level calculations (e.g., binding energies and distances for small model amines) and/or perform a sensitivity analysis over the Hg2+ epsilon and sigma parameters; the capacity claim should be shown to be robust under that uncertainty.","section":"Section 4 and Table 2"},{"comment":"The DFT validation is not independent of the MD force field: the authors 'selected several typical configurations ... obtained from our MD simulations and used them as input for structure optimization using DFT.' This confirms that the MD-generated coordination topologies are local minima at the DFT level, but it cannot establish that the MD sampling produced the correct Hg-amine interaction or that the four-ion complex is the thermodynamically preferred state. The authors should also run DFT optimizations starting from alternative, non-MD initial geometries (e.g., unbound ions, different coordination numbers) and compare relative energies, to test whether the four-ion complex is preferred over other stoichiometries.","section":"Section 5"},{"comment":"All coordination numbers and RDFs are derived from a single 4 ns production trajectory per system (trajectory stored at 1 ps), with no replicate simulations, block averaging, or error bars. Quantities like 'average Hg-N coordination number is 2.5' for PEI-10 + 2 Hg2+ are therefore presented without statistical uncertainty, and the conclusion that five ions are unstable rests on one short run. Please report statistical uncertainties and, ideally, multiple independent initial configurations; this is essential for the capacity claim because coordination numbers directly determine the maximum ion load.","section":"Section 4 and Stage 4"}],"minor_comments":[{"comment":"The column header 'MD Binding enegry, MD' contains a typo ('enegry') and a duplicated 'MD'; the table would be clearer with explicit level-of-theory labels in each column header.","section":"Section 5, Table 3"},{"comment":"The paper does not explicitly state the protonation state of PEI; the assigned charges correspond to neutral amines, but PEI chelation is pH-dependent. This assumption should be stated and justified in the context of water-treatment conditions.","section":"Section 2"},{"comment":"'Data will be made available on request' is not sufficient for a computational study of this type; depositing input topologies, force-field parameter files, and starting coordinates would substantially improve reproducibility.","section":"Data Availability"},{"comment":"The alignment procedure used to superimpose MD and DFT structures is not described, and the color code in the captions is not explained in the main text; a sentence in Section 5 describing the best-fit alignment would help.","section":"Figures 11 and 12"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is within the scope of the journal and I see no evidence of inappropriate citation or circular reasoning. The main risk is that the headline capacity number will be cited without the force-field caveat; the decision should hinge on whether the authors can provide the requested benchmarking and sampling evidence, not on novelty."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper is a straightforward MD + DFT study of Hg2+ binding to short linear PEI chains. The new, usable result is quantitative: a PEI-10 chain coordinates up to four Hg2+ ions, with coordination numbers per ion dropping from about three to two as the ion count rises. That is a genuinely new data point for the design of PEI-based sorbents, and the authors are careful to test several chain lengths and ion loads. They also cross-check their MD structures with M06-2X DFT optimizations and report adsorption energies. So there is real work here.\n\nThe soft spot is the force field. The Hg2+ Lennard-Jones parameters come from Babu-Lim, fitted to Hg2+ hydration, and the cross terms use the geometric combination rule. There is no benchmark against known Hg-amine complexes or binding energies. More tellingly, the paper contains its own internal warning: the MD Hg-N first peak is around 2.15 Å, while the DFT optimized distances in Table 2 are all about 2.5 Å. That is a 0.35 Å offset in the first coordination shell. At 2.15 Å the Coulomb interaction between Hg2+ and the OPLS amine nitrogen (q=-0.78) is very strong, so the model likely overbinds. The DFT does not rescue this, because the optimizations start from MD-selected conformations; they test local stability of those topologies, not whether the force field produces the correct Hg-amine interaction. So the central four-ion capacity should be treated as conditional until the Hg-N interaction is re-benchmarked.\n\nThere is also the protocol point: Stage 1 pre-forms the complexes in vacuum, so the production runs test whether pre-formed complexes stay bound, not whether they assemble from unbound components. The five-ion failure is the only binodal point, and it rests on a single trajectory. No error bars or multiple seeds are reported. Data availability is \"on request,\" which is weak.\n\nAll that said, the paper is not nonsense. The qualitative picture -- short PEI wraps a single Hg2+ with high coordination, longer chains spread several ions along the backbone -- is likely right, and the DFT support for the observed topologies is meaningful. The authors are transparent about their parameters and do not overclaim beyond the pre-formed-complex stability test. The central claim is plausible under the model, just not yet validated against reality.\n\nFor peer review: yes, send it to a serious referee. The right referee will ask for a benchmark of the Hg-amine interaction (e.g., gas-phase or QM/MM binding energies to small amine models), spontaneous assembly runs, and seeding statistics. If those come back clean, the four-ion result becomes useful design data. As it stands, I'd cite it as a preliminary computational estimate, not a settled capacity.","headline":"Plausible preliminary estimate of PEI-10's Hg2+ capacity, but the unbenchmarked Hg-N force field and a 0.35 Å MD/DFT distance gap keep the four-ion maximum conditional.","tokens_in":15854,"tokens_out":2846,"would_cite":false,"duration_ms":30999,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"A single linear polyethyleneimine chain with ten amino groups chelates up to four Hg2+ ions in water, with molecular dynamics and density functional theory agreeing on the complex structures.","keywords":["mercury chelation","polyethyleneimine","molecular dynamics","OPLS/AA force field","radial distribution function","coordination number","DFT validation","heavy metal removal"],"falsifier":"Measure Hg-N distances and coordination numbers for PEI-Hg2+ complexes by EXAFS or X-ray absorption spectroscopy; if a PEI-10 chain is found to bind five Hg2+ ions, or if the Hg-N distances deviate from the predicted 2.15-2.5 Å range by more than about 0.1 Å, the force-field combination and the four-ion capacity would be falsified. Alternatively, recomputing the capacity with a Hg2+ parameter set fitted to Hg-amine reference data would provide a direct test.","tokens_in":14781,"feed_emoji":"🧪","tokens_out":6717,"duration_ms":63928,"temperature":0.7,"pith_summary":"The paper aims to establish the maximum mercury-chelating capacity of a single linear polyethyleneimine (PEI) chain in water and to describe the structure of the complexes. Using atomistic molecular dynamics with the OPLS/AA force field and SPC/E water, it simulates PEI chains with four, five, and ten nitrogen atoms (PEI-4, PEI-5, PEI-10) together with one to five Hg2+ ions. The central quantitative claim is that PEI-10 can stably coordinate up to four Hg2+ ions, while a fifth ion is driven away by Coulombic repulsion. Density functional theory optimizations of the MD-derived geometries give similar structures and Hg-N distances near 2.5 Å, and computed complexation free energies are negative for all stable complexes. If correct, the result gives a microscopic rule for how many mercury ions a short linear PEI chain can hold, which is the quantity relevant to PEI-based water purification materials.","feed_headline":"One PEI-10 chain grabs up to four mercury ions","feed_subtitle":"Molecular dynamics and DFT agree that a single 10-amine polyethyleneimine chain binds at most four Hg2+ ions in water.","key_machinery":"The load-bearing object is the Hg2+–nitrogen coordination shell, quantified by the Hg-N radial distribution function $g(r)$ and its running coordination number $n(r)$ from the production MD trajectories. The models use the OPLS/AA all-atom force field for PEI, the SPC/E model for water, and Lennard-Jones plus Coulombic interactions for Hg2+, with cross-interactions set by the geometric combination rule. Stability is checked by computing adsorption energies in MD and Gibbs free energies of complexation from DFT geometry optimizations at the M06-2X/LanL2DZ level, with the solvent treated implicitly.","core_discovery":"The paper's central discovery is that a neutral linear PEI molecule with ten amino groups binds up to four Hg2+ ions in aqueous solution, forming stable complexes in which each ion is held by two to five nitrogen atoms of the polymer and completed by water molecules in the first hydration shell. Shorter chains PEI-4 and PEI-5 wrap around a single ion with Hg-N coordination numbers of 3 and 5; PEI-10 uses four nitrogens for one ion and fewer per ion as loading increases. DFT-optimized structures reproduce the MD conformations and give Hg-N distances around 2.5 Å, and the complexation free energy is negative in every stable case. Attempts to load a fifth ion onto PEI-10 did not produce a stable complex, which is why the paper states the chelating capacity as four ions per PEI-10 chain.","pith_inferences":["Extending the paper's result, the four-ion ceiling is a property of a 10-nitrogen neutral chain; longer or branched PEI chains might bind more ions per molecule, but the per-amine ceiling may not scale linearly because Coulombic repulsion between bound ions becomes the limiting factor.","The paper does not benchmark the Hg2+ force field against known Hg-amine complexes; a reparameterization using Hg-N reference data could change the predicted coordination numbers and the four-ion capacity, so the capacity should be read as model-dependent until such a benchmark exists.","Since the simulations use neutral PEI chains, the capacity at low pH with protonated amines is likely lower; testing the same protocol at different protonation states would connect the model to realistic wastewater conditions.","The DFT Hg-N distances of about 2.5 Å are systematically longer than the MD first-peak distances of about 2.15 Å, suggesting the MD force field underestimates Hg-N contact length; comparing against EXAFS could decide which is closer to reality."],"forward_implications":["A single PEI-10 chain saturates at four Hg2+ ions; a fifth ion is not stably bound because Hg-Hg Coulombic repulsion wins.","For one ion, shorter chains wrap more completely: PEI-5 coordinates five nitrogens and binds more strongly (MD binding energy -670 ± 21 kJ/mol) than PEI-4 or PEI-10.","Each additional Hg2+ on PEI-10 lowers the per-ion binding strength, so the Gibbs free energy per ion becomes less negative, from -61 kJ/mol for one ion to -21 kJ/mol for four ions at the mixed-basis DFT level.","In multi-ion complexes the chain stretches: Hg-Hg minimum distances are about 5.55 Å for four ions and larger for fewer ions, and each ion keeps only two nitrogen contacts.","DFT-optimized structures agree qualitatively with MD for all stable complexes, supporting the use of this MD protocol to screen PEI architectures for mercury binding."],"supporting_citations":[{"why":"Supplies the Hg2+ Lennard-Jones parameters used for the ion in all simulations.","marker":"[36]"},{"why":"Provides the precedent of applying these Hg2+ parameters with OPLS/AA and SPC/E for Hg2+ removal simulations.","marker":"[33]"},{"why":"Applies the same mercury ion force-field parameters in functionalized nanotube systems, supporting transferability.","marker":"[35]"},{"why":"Defines the OPLS/AA force field used for all PEI bonded and non-bonded parameters.","marker":"[29]"},{"why":"Defines the rigid SPC/E water model and the SHAKE constraints used for the aqueous environment.","marker":"[39]"},{"why":"Supplies the M06-2X density functional used for DFT geometry optimizations.","marker":"[41]"},{"why":"Supplies the LanL2DZ effective-core-potential basis set used for mercury in the DFT validation.","marker":"[42]"}],"fun_headline_variants":["PEI-10 chelates four Hg2+ ions in water","Single PEI-10 chain binds up to four mercury ions","Four Hg2+ ions fit on one PEI-10 chain","MD shows PEI-10 wraps four Hg2+ ions","Ten-amine polymer chelates up to four mercury cations"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole capacity result rests on the assumption that the Hg2+ Lennard-Jones parameters fitted to mercury hydration, combined with the OPLS/AA PEI parameters through the geometric combination rule, correctly capture Hg-amine binding; the paper never tests this combination against known Hg-amine structures or binding energies.","fun_headline_variants_meta":{"raw":{"variants":["PEI-10 chelates four Hg2+ ions in water","Single PEI-10 chain binds up to four mercury ions","Four Hg2+ ions fit on one PEI-10 chain","MD shows PEI-10 wraps four Hg2+ ions","Ten-amine polymer chelates up to four mercury cations"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000565,"raw_usage":{"total_tokens":2634,"prompt_tokens":855,"completion_tokens":1779,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":471,"completion_tokens_details":{"reasoning_tokens":1692}},"tokens_in":471,"tokens_out":1779,"duration_ms":14370,"temperature":1.0,"reasoning_tokens":1692,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T23:14:51.954819+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure Hg-N distances and coordination numbers for PEI-Hg2+ complexes by EXAFS or X-ray absorption spectroscopy; if a PEI-10 chain is found to bind five Hg2+ ions, or if the Hg-N distances deviate from the predicted 2.15-2.5 Å range by more than about 0.1 Å, the force-field combination and the four-ion capacity would be falsified. Alternatively, recomputing the capacity with a Hg2+ parameter set fitted to Hg-amine reference data would provide a direct test.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the Hg2+ Lennard-Jones parameters used for the ion in all simulations."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the precedent of applying these Hg2+ parameters with OPLS/AA and SPC/E for Hg2+ removal simulations."},{"cited_title":"Anitha, S","cited_arxiv_id":null,"evidence_quote":"Applies the same mercury ion force-field parameters in functionalized nanotube systems, supporting transferability."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the M06-2X density functional used for DFT geometry optimizations."}],"review_version":1}