{"id":"0d3ef391-dc22-402f-934d-bb9896702ec1","arxiv_id":"2607.15514","paper_version":2,"verdict":"REJECT","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"high","formal_verification":"none","parameter_count":0,"one_line_summary":"DFT predicts 2D BP3 and BAs3 monolayers store Li (up to 775 mAh/g) and Na (up to 3875 mAh/g) with low diffusion barriers, but the sodium multi-layer capacity claim is likely overstated.","lead":"Using density functional theory, the authors predict that 2D BP3 and BAs3 monolayers could serve as high-capacity, fast-charging battery anodes for lithium and sodium ions, with very low diffusion barriers. The lithium numbers are plausible, but the headline sodium capacity of 3875 mAh/g comes from treating multilayer sodium metal as part of the anode, which needs much stronger evidence.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Maximum sodiation stoichiometry is internally inconsistent: Fig. 6 and the Conclusions show Na72B8X24, which is x=9, not the claimed Na15BX3/Na12BX3 (x=15/12) used to obtain 3875/1365 mAh/g.","rationale":"The reader's verdict is REJECT, and I agree that the paper should be rejected in its current form. The most load-bearing problem is not the generic PBE error bar on the Na-metal reference, although that is also valid, but a direct internal contradiction: the stoichiometry used to compute the headline SIB capacities does not match the stoichiometry shown in the paper's own maximum-sodiation structures. The supercell is B8X24; Na72B8X24 is x=9 per formula unit, giving capacities of 2325/1024 mAh/g, not 3875/1365. The claimed values require Na120B8P24 and Na96B8As24, which never appear in any figure or table. This makes the paper's strongest quantitative claim unsupported by its own data, independent of any DFT functional debates. The reader's identified weakest assumption (multilayer Na thermodynamic stability vs. plating) is related and valid, but the stoichiometry inconsistency is sharper and more decisive: even if the multilayer phase is thermodynamically favored, the capacity number is wrong if the calculated structure has x=9. I therefore set verdict_should_be to REJECT and mark agreement as 'partial' because the reader's phrasing focused on PBE/vdW/AIMD stability rather than the arithmetic mismatch, though both point to the same SIB overclaim.","tokens_in":15826,"tokens_out":5894,"duration_ms":66535,"concrete_test":"Inspect the maximum-sodiation input/output structures (or rerun the reported calculation) and count the actual number of Na atoms in the B8X24 supercell. If the structure is Na72B8X24, recompute the capacity with x=9: BP3 gives 2325 mAh/g and BAs3 gives 1024 mAh/g, directly contradicting the abstract. If the structure is instead Na120B8P24 and Na96B8As24, then the capacity arithmetic is internally consistent and the remaining question is whether PBE+vdW or AIMD confirms that these multilayer phases are more stable than bulk Na — but the first check settles whether the claimed 3875/1365 values correspond to any calculated structure.","verdict_should_be":"REJECT","load_bearing_attack":"The SIB capacity numbers are the centerpiece of the paper's claim of 'ultrahigh' capacity. Equation 3 computes capacity from x_max, and the text states x_max = 15.0 for Na on BP3 and 12.0 for Na on BAs3, giving 3875 and 1365 mAh/g respectively. But the DFT supercell is explicitly B8X24 (8 B, 24 X). Na15BX3 per formula unit corresponds to Na120B8X24; Na12BX3 corresponds to Na96B8X24. The paper's own Fig. 6 and Conclusions, however, describe the 'heavily sodiated' maximum-storage structures as Na72B8X24. That supercell contains 72 Na atoms, i.e. 9 Na per BX3 formula unit, not 15 or 12. Under the paper's capacity formula, x=9 would give 9×26801/103.72 = 2325 mAh/g for BP3 — exactly the value of the earlier BP3 SIB study [40] — and 9×26801/235.57 = 1024 mAh/g for BAs3, not 3875/1365. This is not a subtle DFT-error or vdW issue; it is an arithmetic contradiction in the reported central result. The definition of x in Eq. 1 compounds the problem: the text says N_X=8 for the host supercell while also saying the supercell contains 24 pnictogen atoms, so the denominator is off by a factor of 3. Unless the authors have separate, unreported calculations for Na120B8P24 and Na96B8As24, the 'outstanding' SIB capacity claim is unsupported by the data shown. The Li-ion capacity (x=3, 775/341 mAh/g) is internally consistent and may survive, but the dual-ion claim collapses.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript uses first-principles DFT (PBE, ONCV pseudopotentials) to evaluate two-dimensional hexagonal BP3 and BAs3 monolayers as dual-ion (Li and Na) battery anodes. It reports adsorption-site energetics, CI-NEB diffusion barriers, formation-energy convex hulls, OCV profiles, specific capacities, Bader/COHP bonding analyses, and density-of-states data. The central claims are: Li storage up to Li3BX3 with capacities 775 mAh/g (BP3) and 341 mAh/g (BAs3) at ~0.35-0.39 V; Na storage via multilayer metallic clustering up to Na15BP3/Na12BAs3 with ultrahigh capacities 3875/1365 mAh/g at ~0.15-0.18 V; low diffusion barriers (0.19-0.40 eV); and metallic conductivity maintained throughout charging. The Li results are internally consistent and survive scrutiny; the Na 'ultrahigh-capacity' claim contains an arithmetic inconsistency that is load-bearing.","tokens_in":16240,"tokens_out":5838,"duration_ms":59363,"significance":"If the reported Na-storage stoichiometry and capacities were reliable, the work would present a striking SIB anode candidate with capacities far above most 2D materials. The paper does include systematic calculations (site screening, convex hull, CI-NEB, Bader/COHP, DOS) and the Li-ion capacities are plausible and competitive. However, the marquee result—the Na capacity of 3875/1365 mAh/g—rests on an inconsistent definition of the maximum sodiation stoichiometry. The manuscript's own Figures and Conclusions identify the heavily sodiated supercell as Na72B8X24, which corresponds to x=9 per formula unit, not the x=15/12 used in the capacity equation. If the actual maximum is x=9, the capacities reduce to 2325/1024 mAh/g, which are far less distinctive. Because the central SIB claim is not supported by the data as presented, the significance of the work hinges on whether the authors can supply the missing high-concentration calculations or revise the claims accordingly.","major_comments":[{"comment":"The central SIB capacity claim is internally inconsistent. The capacity in Eq. (3) is computed from x_max=15.0 (BP3) and 12.0 (BAs3), giving 3875 and 1365 mAh/g. However, the only explicitly identified heavily sodiated supercell in the manuscript is Na72B8X24 (Fig. 6 and Conclusions), which corresponds to x=9 per formula unit, not the claimed values. Under Eq. (3), x=9 gives 2325 mAh/g for BP3 and 1024 mAh/g for BAs3—the former exactly matching the prior BP3 SIB study cited as Ref. [40]. The manuscript does not report structures, total energies, or formation energies for Na120B8P24 or Na96B8As24 anywhere, although Table I lists rows at x=12, 13.5, and 15.0. Either the Table I entries at x>9 are based on unshown calculations, in which case the data must be presented, or the quoted maximum stoichiometry and capacities are wrong. As written, the 'ultrahigh' SIB capacity claim is unsupported","section":"Eq. (3), Table I, Fig. 6, Conclusions"},{"comment":"The definition of x is ambiguous and appears to contain a factor-of-3 error. Eq. (1) and the text define x = N_adatom/N_X with N_X=8 for the 'host supercell.' The supercell is explicitly stated to be B8X24, containing 24 pnictogen atoms, so N_X should be 24 if X denotes pnictogen atoms. The text later states 'x=1/8 (0.125) corresponds to a single adsorbed alkali atom,' which is consistent with x being defined per BX3 formula unit (8 formula units in the cell) rather than per pnictogen atom. This mislabeling does not affect the numerical results if the intended quantity is per formula unit, but it creates confusion in every concentration label and should be corrected unambiguously.","section":"Eq. (1), Section III.A"},{"comment":"The thermodynamic stability of multilayer Na phases at x>3 is asserted solely from PBE formation energies relative to bcc Na. The energy gain at x=15 is only -0.18 eV/atom for BP3 and -0.13 eV/atom for BAs3 in Table I. This margin is comparable to PBE's systematic errors for alkali-metal cohesive energies and well within the range where missing van der Waals interactions or vibrational contributions could change the sign. The paper provides no AIMD, phonon, or vdW-corrected checks for these multilayer structures, nor does it consider competition with Na-X compound formation. The positive OCV plateaus near 0.10 V and 0.03 V at high x are correspondingly fragile. The authors should either provide more robust evidence for the multilayer phases (e.g., vdW-DFT calculations, AIMD trajectories, or at least a discussion of the error bar) or moderate the 'safe, dendrite-free' claim.","section":"Section III.A, Fig. 1c, Table I"}],"minor_comments":[{"comment":"The caption states the visible sodiation structures extend up to x=10.50, while Table I reports values up to x=15.0. If structures for x>10.5 were calculated, they are not shown; if they were not, Table I is unexplained.","section":"Fig. 2 caption"},{"comment":"The phrase 'guaranteeing a safe, dendrite-free operating mechanism near 0 V' is too strong. Even if the OCV remains positive, the margin of ~0.02-0.10 V at high x is small against typical DFT errors, and safety claims from thermodynamic hull data alone are speculative.","section":"Section III.A"},{"comment":"The intrinsic phase stability of the pristine BP3/BAs3 monolayers is justified by citing two unpublished arXiv preprints by the same authors. Since these are not peer-reviewed, the stability premise is not independently established in the present manuscript.","section":"References [41,42]"},{"comment":"The abstract calls the materials 'structurally resilient,' but the manuscript contains no mechanical, thermal, or dynamic stability analysis beyond the PBE adsorption calculations. Either provide such evidence or soften the wording.","section":"Abstract/Conclusions"}],"recommendation":"major_revision","confidential_remarks":"The inconsistency between the claimed Na15BX3/Na12BX3 stoichiometry and the reported Na72B8X24 structures is the crux. If the authors cannot provide the missing calculations for x=15/12, the SIB capacity claims must be revised downward to x=9, which would erase the paper's main novelty and likely reduce it to an incremental study. I recommend requesting the raw data for all concentration points and a careful re-derivation of the capacity values before reconsidering the manuscript."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe headline SIB capacities (3875 and 1365 mAh/g) do not survive contact with the paper's own numbers. Eq. (1) defines x as adatoms per pnictogen, with N_X = 8 for the \"host supercell,\" but the supercell is explicitly B8X24 — 8 boron and 24 pnictogens. Setting x = 1/8 for one adsorbate works only if x is per formula unit (per B), so the definitional text is wrong. More damaging, the maximum sodiation structure shown in Fig. 6 and quoted in the Conclusions is Na72B8X24, which is x = 9 per BX3, not the x = 15 and x = 12 used in Table I and Eq. (3) to get 3875 and 1365 mAh/g. At x = 9, BP3 gives 2325 mAh/g — exactly the prior BP3 SIB value [40] — and BAs3 gives 1024 mAh/g, not 1365. That is an arithmetic contradiction in the central claim.\n\nWhat is actually new and decent: the Li results. The lithiation convex hull is internally consistent, Li3BX3 is a sensible terminal phase, and 775/341 mAh/g follow from the stated formula. The CI-NEB barriers (0.19–0.40 eV) are plausible and the COHP/PDOS analysis is routine but executed cleanly. The comparison with the earlier Na-on-BP3 study [40] is honest, and BAs3 is a legitimate extension.\n\nThe soft spot beyond the stoichiometry error is the multilayer Na model itself. The stability of Na15BX3 (or Na9BX3, whichever they actually computed) rests on PBE formation energies relative to bcc Na with a margin of only about −0.18 eV/atom at the top. No vdW correction, no AIMD, and that margin is within PBE's typical error for alkali-metal cohesion. So the \"safe, dendrite-free\" sodiation near 0 V is not established. This is load-bearing: the abstract's \"ultrahigh capacities\" rest entirely on it.\n\nThe reader's take matches mine. I'd also note that Fig. 2 shows Na up to x = 10.5, so there are at least three different maximum concentrations in the paper (15, 12, 9, and 10.5 depending on the figure). That kind of inconsistency makes it impossible to trust the reported SIB numbers.\n\nWho gets value from this? Someone compiling a table of 2D anode candidates might mine the Li half, but the SIB claims should not be cited. I'd use it in a reading group as a cautionary example of why you check the stoichiometry in capacity formulas before trusting the abstract.\n\nRecommendation: desk reject in current form. The Li portion might be salvageable in a serious revision, but only if the authors fix the sodiation concentration, do vdW and AIMD checks, and correct or drop the SIB capacity claims.","headline":"Routine DFT screening with a plausible Li result; the SIB capacity claim is undone by an internal stoichiometry contradiction.","tokens_in":16767,"tokens_out":6121,"would_cite":false,"duration_ms":54628,"reading_group":"maybe","serious_thinker":"no","would_accept_peer_review":false},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Hexagonal BP3 and BAs3 monolayers are shown to be fast-charging, high-capacity anodes for both lithium- and sodium-ion batteries, staying metallic at every state of charge.","keywords":["two-dimensional anode","BP3 monolayer","BAs3 monolayer","lithium-ion battery","sodium-ion battery","density functional theory","diffusion barrier","multilayer sodiation"],"falsifier":"Recompute the formation energies of NaxBX3 with dispersion corrections and run ab initio molecular dynamics at 300 K for Na15BP3. If the Na15 phase decomposes into bulk sodium plus lower-concentration phases, or if its voltage plateau departs from the predicted ~0.10 V, the ultrahigh-capacity claim would be falsified; an experimental measurement of the sodium plating/stripping potential on BP3 would settle it directly.","tokens_in":15657,"feed_emoji":"⚡","tokens_out":7116,"duration_ms":65439,"temperature":0.7,"pith_summary":"The paper sets out to show that two atomically thin materials — hexagonal monolayers of boron phosphide (BP3) and boron arsenide (BAs3) — can serve as dual-use battery anodes, storing lithium and sodium while remaining electrically conductive. Using density functional theory, the authors find that both hosts bind alkali ions most strongly at the same hollow site, and that ions hop between these sites with small energy barriers (0.19–0.40 eV), which implies fast charging. They report that lithium saturates at Li3BX3, giving BP3 a capacity of 775 mAh/g — about double graphite — while sodium storage extends to multilayer clusters, reaching 3875 mAh/g on BP3 at a low average voltage of 0.18 V. If these predictions hold, the materials would combine high capacity, low operating potential, and metallicity without conductive additives.","feed_headline":"Sodium storage up to 3875 mAh/g on a 2D boron-phosphide sheet","feed_subtitle":"DFT study also predicts fast Li/Na diffusion (0.19–0.40 eV) and metallic conductivity in BP3/BAs3 monolayers.","key_machinery":"The central object is the 2D hexagonal BX3 monolayer with a puckered lattice (X = P or As). The argument runs on three mechanisms: (1) an adsorption hierarchy that puts the H3 hollow site deepest, driven by combined ionic charge donation and covalent p-p hybridization; (2) a direct H3-to-H3 hopping path that skips a high-energy intermediate site, giving low diffusion barriers; and (3) for sodium, a transition beyond monolayer coverage into multilayer metallic clustering that keeps voltages positive and extends storage to extreme concentrations. These are linked through formation-energy convex hulls and open-circuit voltage profiles that identify the stable intercalation phases.","core_discovery":"The central claim is that 2D hexagonal BX3 (X = P, As) monolayers are high-performance dual-ion anodes: they bind Li and Na strongly at the H3 hollow site, diffuse ions quickly via direct H3-to-H3 hops (0.19–0.40 eV barriers), and stay metallic at every state of charge. Lithium saturates at Li3BX3, giving 775 mAh/g (BP3) and 341 mAh/g (BAs3) at average voltages of 0.39 and 0.35 V. Sodium continues to multilayer clusters up to Na15BX3 (Na12BX3 for BAs3), producing ultrahigh capacities of 3875 and 1365 mAh/g at 0.18 and 0.15 V. The stability of the multilayer sodiated phases is what makes the extreme sodium capacities possible.","pith_inferences":["The same multilayer-storage mechanism, if stable, may extend to potassium or other alkali ions on these hosts, potentially yielding similar ultrahigh capacities.","Real devices would need to verify volumetric capacity and whether the multilayer sodium clusters survive finite temperature and cycling; the current evidence is zero-temperature and dispersion-free.","The H3-site preference and direct hopping path suggest that strain engineering or chemical substitution on the pnictogen sublattice could tune diffusion barriers and capacities.","Because the paper identifies formation energies relative to bulk Na, a direct experimental test of sodium plating potential on BP3 would cleanly validate or overturn the ultrahigh-capacity claim."],"forward_implications":["A BP3 anode could store twice as much lithium per gram as commercial graphite (775 vs 372 mAh/g) while charging at rates comparable to or better than graphite.","Sodium-ion batteries built on BP3 could reach gravimetric capacities near 3875 mAh/g, far above typical MXenes and most 2D anodes, if the multilayer sodiation phase is real.","Both hosts stay metallic throughout charge/discharge, so the electrode may not need conductive carbon additives, reducing weight and cost.","The low diffusion barriers (0.19–0.40 eV) imply fast room-temperature ion transport, which supports rapid charging.","Low average operating voltages (0.15–0.39 V) help maximize full-cell energy density."],"fun_headline_variants":["2D BP3 monolayer stores sodium at 3875 mAh/g capacity","BAs3 and BP3 anodes: low-barrier Li/Na diffusion, metallic throughout","Sodium capacity of 3875 mAh/g on 2D BP3 via multilayer storage","Fast Li/Na anodes: 2D BX3 sheets, barriers as low as 0.19 eV"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The ultrahigh sodium capacities rest on the assumption that multilayer sodium clusters up to Na15BX3 are thermodynamically stable against separation into bulk sodium, with a computed margin of only about 0.18 eV per atom using a functional that is known to be imperfect for alkali-metal cohesion and without van der Waals or finite-temperature checks.","fun_headline_variants_meta":{"raw":{"variants":["2D BP3 monolayer stores sodium at 3875 mAh/g capacity","BAs3 and BP3 anodes: low-barrier Li/Na diffusion, metallic throughout","Sodium capacity of 3875 mAh/g on 2D BP3 via multilayer storage","Fast Li/Na anodes: 2D BX3 sheets, barriers as low as 0.19 eV"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000342,"raw_usage":{"total_tokens":1832,"prompt_tokens":970,"completion_tokens":862,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":714,"completion_tokens_details":{"reasoning_tokens":762}},"tokens_in":714,"tokens_out":862,"duration_ms":8703,"temperature":1.0,"reasoning_tokens":762,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T23:05:24.879834+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Recompute the formation energies of NaxBX3 with dispersion corrections and run ab initio molecular dynamics at 300 K for Na15BP3. If the Na15 phase decomposes into bulk sodium plus lower-concentration phases, or if its voltage plateau departs from the predicted ~0.10 V, the ultrahigh-capacity claim would be falsified; an experimental measurement of the sodium plating/stripping potential on BP3 would settle it directly.","supporting_citations":[],"review_version":1}