{"id":"10be2179-ee58-45af-9bdd-21e57036619f","arxiv_id":"2608.12439","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Strain mismatch between asymmetric surface terminations drives Janus MXene sheets to self-scroll, and the resulting nanoscroll can wrap a nanoparticle while locally widening its interlayer spacing.","lead":"Using molecular dynamics simulations, this paper shows how Janus MXene sheets with different chemical endings on their two faces roll themselves into nanoscrolls, driven by strain between the faces. It also shows that a nanoparticle can be wrapped inside the scroll, widening the layers, but hydrogen gas released during wrapping may form bubbles that hurt batteries.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The core mechanism depends on ReaxFF predictions for -O/-OH terminated Janus MXenes, but validation (Table 1) covers only bare Ti2C; without DFT checks of termination energetics and strain mismatch, the scroll-driving bending moments and H2 release are unsupported.","rationale":"The manuscript presents a plausible MD-based mechanism for Janus MXene scrolling, but its central quantitative ingredient is the surface strain mismatch in Table 2, which is directly used to derive bending moments (Eqs. 3-5) and to predict scroll geometry, interlayer spacing, and encapsulation behavior. All of these quantities originate from ReaxFF simulations. The only validation in Table 1 is for bare Ti2C: lattice constant, thickness, Ti-C bond, and Young's moduli, with no test of -O/-OH terminated structures or of the reactive chemistry (O-H dissociation, H2 formation) underpinning the encapsulation claim and the battery-safety caveat. Since ReaxFF parameters are known to be system-specific, transferability to functionalized Janus MXenes is a genuine load-bearing assumption. The paper's own Limitations section (3.6) addresses only the choice of TiNP as a model particle, not the missing validation of the force field for the functionalized surfaces. A DFT benchmark of the strain mismatch and reaction energetics is therefore the decisive check. If that check showed significant deviation, the predicted scroll geometries and the H2 release mechanism would not be reliable. The reader's conditional verdict is appropriate: the mechanism is plausible but not yet established. I agree with the reader's weakest_assumption and found no stronger internal inconsistency that would move the verdict to reject or unverified.","tokens_in":21980,"tokens_out":6173,"duration_ms":53206,"concrete_test":"Run PBE-D3 (or SCAN) DFT relaxations of the three Janus Ti2C supercells used in Classes I-III (bTi2C(O), bTi2C(OH), (O)Ti2C(OH)), extract the relaxed Ti-Ti bond lengths on each terminated surface, and compare the resulting strain mismatch with Table 2. If the DFT strain mismatch differs by more than ~1 percentage point or changes sign, the ReaxFF-based bending-moment mechanism and the predicted scroll morphologies are not transferable.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Table 1 validates ReaxFF [38] only against bare Ti2C lattice constant, thickness, Ti-C bond length, and directional Young's moduli (the latter with errors up to 12.83%). The manuscript nevertheless uses this force field to simulate -O and -OH terminated Janus MXenes, including O-H dissociation and H2 formation during TiNP encapsulation. The strain values in Table 2, which form the quantitative basis for the bending-moment model in Eqs. (3)-(5), are extracted from ReaxFF-relaxed Janus structures. If ReaxFF misrepresents the relative stability, bond lengths, or surface energetics of -O versus -OH terminations, the reported strain mismatches (e.g., -1.6% vs +3.6% for (O)Ti2C(OH)) and the resulting scroll geometries, interlayer distances, and scrolling directions could be artifacts. No DFT benchmark is provided for any functionalized MXene, and the cited more-detailed validation [44] is a self-cited preprint rather than independent work. The ReaxFF parameters [38] were developed for MXene intercalation and water dynamics, not specifically calibrated for O-H dissociation on MXene surfaces; the H2 release mechanism in Section 3.5 relies on O-H bond-breaking and Ti-O bond-formation energetics that are not validated. Because the central claim is that lattice-induced strain generates the bending moment driving scrolling, the entire causal story rests on the force field's fidelity for these functionalized surfaces. The authors' Limitations section (3.6) acknowledges the TiNP is a model for Si/S but does not flag this missing functionalization validation.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports reactive molecular dynamics simulations of three Janus Ti2C MXene surface terminations (bare, -O, and -OH) under four boundary-condition classes, and claims that asymmetric surface termination produces 1-7% lattice-induced strain, which creates a bending moment about carbon pivot atoms and drives spontaneous curling, scrolling, or nanotube formation depending on flake size. It further reports that a Ti nanoparticle can be encapsulated by a scrolling sheet, locally enlarging the interlayer spacing, and that encapsulation is accompanied by H2 release. The paper validates the ReaxFF potential against DFT/experimental structural and elastic data for bare Ti2C, then uses an analytical-MD bending-moment model to explain scrolling directions and to compare moments across terminations.","tokens_in":22301,"tokens_out":5943,"duration_ms":52177,"significance":"The central qualitative observation, namely that flat Janus MXene sheets scroll spontaneously in MD without imposed moments, is directly demonstrated and is a useful step toward understanding the experimentally reported scalable synthesis route for MXene scrolls. The systematic morphology map as a function of surface termination and flake size, and the identification of H2 as a side product during nanoparticle encapsulation, are valuable and falsifiable predictions. However, the quantitative causal chain (strain mismatch, bending moment, scroll geometry) rests on force-field transferability to functionalized surfaces and on strain measurements made after relaxation on curved structures; these supports are currently weaker than the conclusions require. If the requested benchmarks and reanalyses are provided, the paper will make a solid contribution.","major_comments":[{"comment":"The force-field validation is confined to bare Ti2C: Table 1 compares ReaxFF predictions against DFT for the in-plane lattice parameter, thickness, Ti-C bond length, and Young's moduli, with systematic errors up to 12.83% in the moduli. The simulations then use the same ReaxFF potential to model -O and -OH terminated surfaces, O-H dissociation, and H2 formation in Section 3.5, with no DFT benchmark for any functionalized structure. Because the strain mismatch in Table 2 and the H2 release mechanism both depend on the energetics and bond-length changes of these functional groups, the quantitative scrolling mechanism is unsupported unless the potential is validated for the functionalized systems. I request DFT calculations of equilibrium bond lengths, relative stabilities, and O-H dissociation energetics for the relevant terminations, or a clear statement that the quantitative claims are conditional on force-field fidelity.","section":"Section 3.1 (Table 1)"},{"comment":"The strain values that drive the bending-moment model are measured after energy minimization on already-curled sheets, \"near the end regions\" (Table 2). These post-relaxation strains combine the intrinsic lattice mismatch with the geometric strain of curvature, and they are then substituted into Eqs. (3)-(5) as the pre-scrolling mismatch that generates the bending moment. This is a circular step for the quantitative claim. I note that the MD observation of spontaneous scrolling from flat sheets is not itself circular; however, the authors should compute the strain from flat, uncurled reference structures (e.g., before minimization or with curvature constrained) and confirm that the values in Table 2, and the ordering of bending moments, do not depend on the local curvature at the measurement site.","section":"Section 3.2 (Table 2) and Section 3.5 (Eqs. (3)-(5))"},{"comment":"The analytical bending-moment calculation contains an ad hoc geometric parameter: the nominal cross-sectional area is defined with r approximately 2.5 Angstrom, \"chosen to be slightly larger than the Ti-C bond length,\" and the position vectors are set equal to the Ti-C bond length. The absolute bending moments (2.18x10^-19 to 4.46x10^-19 N.m, and 5.39x10^-19 and 6.64x10^-19 N.m) therefore depend on this choice, and the claim that MD and analytical-MD agree \"with the same order of magnitude\" is not a strong consistency check unless the sensitivity to r is quantified. Please provide a sensitivity analysis over a physically plausible range of r and show that the qualitative ordering of the bending moments, which is the basis for the scrolling-direction and morphology claims, is robust.","section":"Section 3.5, Eqs. (1)-(5)"},{"comment":"The H2-release prediction is a central applied finding, but it is supported only by a qualitative electronegativity argument and by reactive trajectories from a force field not validated for O-H dissociation on MXene surfaces. No DFT barriers or experimental benchmark are provided. Because the bubble-shielding claim is one of the paper's main practical conclusions, this mechanism should be corroborated by independent electronic-structure calculations or at least reported with an explicit uncertainty assessment.","section":"Section 3.5 and Section 3.6"}],"minor_comments":[{"comment":"The x and y crystallographic directions are described inconsistently: Section 2.1 calls x:[100] the zigzag direction and y:[010] the armchair direction, whereas the Figure 1 caption says x is armchair and y is zigzag. Please correct this.","section":"Section 2.1 and Figure 1 caption"},{"comment":"The enumeration of simulation classes lists \"Class I, II, III, and VI,\" but the fourth class is Class IV. Please correct the typo.","section":"Section 2.1"},{"comment":"The surname of the corresponding author appears as \"Darban\" in the author list and \"Darband\" in the Author Contributions section; please standardize.","section":"Author list and Author Contributions"},{"comment":"The declaration reads \"The author declares\" but the manuscript has multiple authors; this should be plural.","section":"Declaration of Interests"},{"comment":"There are repeated spelling errors, e.g., \"nanoscrrolled\" instead of \"nanoscrolled\" in several figure-related passages.","section":"Figures 2, 3, 8 and text"},{"comment":"The equations are poorly typeset and some symbols are undefined or ambiguous (for example, the nominal radius r and the cross-sectional area A, and the superscripts/subscripts on the moment terms). Please define all symbols in one consistent notation.","section":"Equations (1)-(5)"},{"comment":"The text states that simulations were run for at least 75 ns, but the reported energy and MSD plots show only the first 1 ns (and Figure 8 refers to a 2 ns snapshot). Please clarify the displayed time interval and whether the longer runs change the reported morphologies.","section":"Section 2.2 and Figures 2-5, 8"}],"recommendation":"major_revision","confidential_remarks":"The decisive issue is the force-field validation gap: the central quantitative claims depend on ReaxFF predictions for functionalized surfaces and for O-H dissociation, while Table 1 validates the potential only against bare Ti2C. I would not recommend acceptance without additional DFT benchmarks for the functionalized systems and a reanalysis of the strains from flat reference geometries. I also note that the detailed validation cited as reference [44] is a self-archived preprint by one of the authors, so it does not constitute independent support. The paper otherwise fits the journal's scope and the qualitative scrolling observation is sound."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Read it. The genuine new result is that reactive MD shows three Janus Ti2C MXenes spontaneously scroll without any imposed moment, with morphology depending on termination and flake size, and that a Ti nanoparticle gets wrapped with locally widened interlayer spacing. That is a concrete, system-specific claim, and the size/termination map in Table 5 is useful design information. The paper also does the field a service by flagging H2 release during encapsulation; I don't know of another MXene scrolling paper that raises that.\n\nThe soft spots are real but they do not kill the central observation. First, the quantitative story is partially circular: strains are read off already-curled, minimized structures and then plugged into Eqs. (3)-(5) to compute the bending moment that supposedly drove the curling. The MD itself shows the scrolling, so the mechanism is not invented, but the numbers from the analytical model should be treated as illustrative. Second, the force-field transferability is the load-bearing weakness, exactly as the stress-test says. Table 1 validates ReaxFF only against bare Ti2C; the simulations then use it for -O/-OH surfaces, O-H dissociation, and H2 formation. A DFT benchmark for the functionalized surfaces, even just relative energies and Ti-O bond lengths, would be needed before I trust the reported 1-7% strains or the H2 release. The self-cited preprint [44] does not substitute for that. Third, the abstract says \"experimental structural and elastic properties\", but Table 1 compares against DFT, not experiment; that should be fixed. Minor: no deposited data/scripts, no repeat simulations or error bars; at 1 K and single trajectories, the quantitative scroll diameters are less convincing than the qualitative trends.\n\nAll that said, the paper is coherent, candid about several limitations, and the central observation is not contradicted. I'd send it to peer review. The right referee report would ask for (i) DFT checks of termination energetics and strain, (ii) control simulations on symmetric sheets, (iii) a softer abstract. This paper is for computational MXene researchers and experimentalists planning scroll synthesis; experimentalists should treat the H2 warning as a hypothesis until the force field is checked.","headline":"The MD observation of spontaneous Janus MXene scrolling looks real and worth reporting, but the bending-moment model and the H2 chemistry ride on an unvalidated force field.","tokens_in":22842,"tokens_out":2897,"would_cite":false,"duration_ms":26579,"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":"Janus Ti2C MXene sheets roll themselves into nanoscrolls because asymmetric surface terminations create a 1–7% lattice strain mismatch that generates a bending moment around carbon atoms.","keywords":["Janus MXenes","nanoscroll","self-rolling","bending moment","core@shell composite","reactive molecular dynamics","interlayer spacing","H2 evolution"],"falsifier":"Compute the relaxed Ti-O bond lengths on the two faces of a free-standing (O)Ti2C(OH) monolayer with density functional theory. If the -O face's Ti-O bond is not shorter than the -OH face's bond (the paper finds 1.92 Å versus 1.95 Å), the reported 5.2% net strain and the bending moment that drives scrolling vanish, and the proposed mechanism would need revision.","tokens_in":21742,"feed_emoji":"🌀","tokens_out":11909,"duration_ms":98144,"temperature":0.7,"pith_summary":"The paper tries to establish an atomistic mechanism for a phenomenon already achieved experimentally: why a Janus (two-faced) MXene sheet curls itself into a nanoscroll. Using reactive molecular dynamics validated against density functional theory, it shows that when the two faces of a Ti2C sheet carry different chemical groups (bare, -O, or -OH), the Ti-O and Ti-Ti distances on the two faces differ enough to produce a 1–7% lattice-induced strain. That strain mismatch creates a net bending moment about the carbon pivot atoms, so the sheet spontaneously rolls toward the face under compressive strain. The final shape is not arbitrary: for long flakes the sheet forms a multiwalled nanoscroll whose inner diameter and interlayer spacing are set by the termination pair, while small square flakes merely curve or, in one termination, close into a chiral nanotube. The same scrolling wraps a single anchored titanium nanoparticle into a core@shell structure, locally widening the interlayer spacing from 0.96 nm to 2.02 nm, but the simulation also shows that encapsulation releases H2 molecules that can form performance-degrading nanobubbles.","feed_headline":"Asymmetric etching rolls MXene flakes into nanoscrolls","feed_subtitle":"Simulations show surface strain mismatch bends the sheets, and a captured particle widens the interlayer spacing","key_machinery":"The load-bearing object is the net bending moment about each carbon pivot atom. The paper's Eq. (3b) writes it as $\\vec{M}_i = A_{(-O,-OH)}(2.25\\sin\\theta_{-OH}+\\sin\\theta_{-O})(-\\hat{y})$, where $A_{(-O,-OH)} = Y A |\\varepsilon_{-O}| r_{\\mathrm{Ti-C}}$ ranges from $2.18\\times10^{-19}$ to $4.46\\times10^{-19}$ N·m. Each carbon sits between three Ti atoms on the -OH face and three on the -O face, so the strain-induced forces on the two faces are parallel but opposite; because they act on opposite sides of the carbon, their moments add instead of cancel. Since the strain acts only along the x direction, only moments about the y axis survive, and the right-hand rule selects the rolling direction toward the compressively strained face. The machinery also explains the metastability of bTi2C(O): both faces are compressive, with a net strain of only 4.5%, and the Ti-C bond-length variation is insufficient to hold a stable curvature, so Coulombic Ti-O interlayer attraction zips the semi-scrolled structure closed.","core_discovery":"On its own terms, the paper's central discovery is that nanoscroll formation in Janus MXenes is a deterministic strain-driven process, not a stochastic curling artifact. Starting from the experimentally known Ti2C(OH)2 sheet and etching one face to -O or bare, the simulations find that the etched face contracts while the -OH face expands: the Ti-O bond length falls from 1.95 Å to 1.92 Å after H removal, giving net strains of -1.6% to -7.1% on the compressed face and tensile strains up to +4.2% on the opposite face. This differential strain produces a net bending moment around the carbon pivot atoms, expressed by Eq. (3b), with an average magnitude near $3.32\\times10^{-19}$ N·m for (O)Ti2C(OH); the moment points the same way for every carbon atom, so the sheet rolls toward the face under compressive strain. Whether the final object is a curved sheet, a chiral nanotube, or a multiwalled nanoscroll is then set by curvature and flake size: at 120 nm length, (O)Ti2C(OH) and bTi2C(OH) form stable scrolls with interlayer spacings around 0.7 nm and inner diameters of 7–8 nm, whereas bTi2C(O) produces a metastable structure with spacing around 1.7 nm and inner diameter above 20 nm. When a single Ti nanoparticle is anchored on the bare face of bTi2C(OH), the sheet still scrolls, bonds to the particle, and locally expands the interlayer spacing to 2.02 nm; the particle's electron transfer to oxygen weakens O-H bonds and releases H2.","pith_inferences":["If the carbon-pivot bending mechanism transfers to other MXene compositions such as Ti3C2, V2C, or Nb2C, the surface bond lengths alone might predict which termination pairs scroll and in which direction, making Eq. (3b) a design rule for a whole family of materials.","The H2 release pathway is tied to electron transfer from the titanium particle to oxygen; a silicon or sulfur core, which the authors could not simulate because their force field has no parameters for those elements, may not release H2 in the same way, so the nanobubble risk should not be assumed universal for core@shell MXene electrodes.","The locally enlarged spacing around a captured particle suggests that ordered arrays of nanoparticles could imprint a periodic interlayer-spacing pattern into a scroll, offering a lithography-free route to engineered ion channels, though scroll uniformity at high particle density remains untested."],"forward_implications":["For long flakes the termination pair sets the scroll geometry: (O)Ti2C(OH) and bTi2C(OH) form stable multiwalled nanoscrolls with interlayer spacing around 0.7 nm and inner diameter 7–8 nm, while bTi2C(O) forms only a metastable structure with spacing around 1.7 nm and inner diameter above 20 nm.","Flake size is a second control: square flakes of 10–20 nm only curve (one 20 nm bTi2C(OH) flake closes into a chiral nanotube), whereas 40–100 nm flakes of the -OH-bearing systems scroll.","Clamping one end of the flake, equivalent to effectively doubling its length, increases the inner diameter by roughly 8–34% and interlayer spacing by up to 46%, depending on termination.","A nanoparticle does not stop scrolling: it is captured between MXene layers, bonds to the sheet, and raises the local interlayer spacing from 0.96 nm to 2.02 nm, so particle size can be a design handle for interlayer channels.","Encapsulation can release H2 that accumulates as nanobubbles, so a practical MXene-scroll electrode must mitigate this bubble-shielding effect."],"supporting_citations":[{"why":"Supplies the reactive force-field model used in every simulation of scrolling and encapsulation.","marker":"[38]"},{"why":"Supplies the DFT reference values for lattice constant, thickness, Ti-C bond length, and Young's moduli used to validate the model.","marker":"[42]"},{"why":"Provides additional DFT structural and elastic data for MXene monolayers used in the validation table.","marker":"[43]"},{"why":"Reports the experimentally verified scalable synthesis of MXene scrolls that this simulation setup is designed to explain.","marker":"[21]"},{"why":"Supplies the Ti2C(OH)2 structure from which the Janus terminations are built by removing H and O atoms.","marker":"[36]"},{"why":"Provides the electronegativity values used to argue that O takes electrons from the Ti particle more readily than H does, driving O-H dissociation and H2 release.","marker":"[58]"},{"why":"Documents spontaneous curling of freestanding Janus transition-metal dichalcogenides, the analogous mechanism this work extends to MXenes.","marker":"[31]"},{"why":"Establishes carbon nanoscroll structure and dynamics, providing the size and aspect-ratio dependence that motivates the square-flake study.","marker":"[35]"},{"why":"Provides additional validation of the force field for titanium-based MXenes.","marker":"[44]"}],"fun_headline_variants":["Strain mismatch rolls MXene sheets into nanoscrolls","Janus MXene sheets curl into nanoscrolls under strain","Particle encapsulation widens nanoscroll interlayer gaps","Simulations reveal strain-driven nanoscroll formation in MXenes","H2 release during nanoparticle encapsulation in MXene scrolls"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument assumes that the reactive force field used for all simulations, which is validated only against bare Ti2C lattice parameters, layer thickness, and Young's moduli, also correctly captures the energetics of -O and -OH termination, O-H bond breaking, and H2 formation on functionalized surfaces.","fun_headline_variants_meta":{"raw":{"variants":["Strain mismatch rolls MXene sheets into nanoscrolls","Janus MXene sheets curl into nanoscrolls under strain","Particle encapsulation widens nanoscroll interlayer gaps","Simulations reveal strain-driven nanoscroll formation in MXenes","H2 release during nanoparticle encapsulation in MXene scrolls"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000767,"raw_usage":{"total_tokens":3550,"prompt_tokens":1248,"completion_tokens":2302,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":864,"completion_tokens_details":{"reasoning_tokens":2217}},"tokens_in":864,"tokens_out":2302,"duration_ms":16879,"temperature":1.0,"reasoning_tokens":2217,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T00:14:54.953017+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the relaxed Ti-O bond lengths on the two faces of a free-standing (O)Ti2C(OH) monolayer with density functional theory. If the -O face's Ti-O bond is not shorter than the -OH face's bond (the paper finds 1.92 Å versus 1.95 Å), the reported 5.2% net strain and the bending moment that drives scrolling vanish, and the proposed mechanism would need revision.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the reactive force-field model used in every simulation of scrolling and encapsulation."},{"cited_title":"Zhang, B","cited_arxiv_id":null,"evidence_quote":"Reports the experimentally verified scalable synthesis of MXene scrolls that this simulation setup is designed to explain."},{"cited_title":"Batsanov, The Concept of Electronegativit y","cited_arxiv_id":null,"evidence_quote":"Provides the electronegativity values used to argue that O takes electrons from the Ti particle more readily than H does, driving O-H dissociation and H2 release."},{"cited_title":"Xiong, J","cited_arxiv_id":null,"evidence_quote":"Documents spontaneous curling of freestanding Janus transition-metal dichalcogenides, the analogous mechanism this work extends to MXenes."},{"cited_title":"Braga, V.R","cited_arxiv_id":null,"evidence_quote":"Establishes carbon nanoscroll structure and dynamics, providing the size and aspect-ratio dependence that motivates the square-flake study."},{"cited_title":"Darban, Auxetic Response in Two-Dimensional MXenes with Atomically Defined Perforations, ArXiv Prepr","cited_arxiv_id":null,"evidence_quote":"Provides additional validation of the force field for titanium-based MXenes."}],"review_version":1}