{"id":"c6209a59-697e-41fb-9a2e-3138c9141ba2","arxiv_id":"2608.07856","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"high","formal_verification":"none","parameter_count":3,"one_line_summary":"Spin-resolved DFT predicts interstitial anionic electron (electride) character in elemental manganese at ambient pressure, based on ELF, charge, and DOS analyses.","lead":"DFT calculations on three phases of elemental manganese find electrons localized in empty spaces of the crystal lattice, which the authors interpret as a 'native electride' property. If correct, the result would add a simple, magnetic element to the electride family, though the evidence is entirely computational and the classification is contested.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The headline IAE charges come from a non-conserving spherical partition and a circular ELF-to-pseudo-atom loop, so the quantitative electride evidence needs a pseudo-atom-free QTAIM test before the central claim can stand.","rationale":"The reader identified the PBE+U model's large lattice mismatch as the weakest assumption, and that is a legitimate concern because a 6–7% overestimate signals an incorrect magnetic order or Hubbard treatment. However, the most load-bearing weakness is more internal: the quantitative electride signature in the abstract is taken from Eq. (3), a spherical-integration scheme that does not tile the cell and therefore does not conserve charge. For α-Mn the unassigned charge is about 29 e, so the -1.645 e per interstitial basin has no rigorous meaning. The Bader analysis does conserve charge, but its basins were created by inserting pseudo-atoms at the exact ELF maxima that the paper uses as evidence of electride character, making the validation circular unless true non-nuclear maxima are demonstrated. A pseudo-atom-free QTAIM/Bader test would settle this directly. My recommendation remains CONDITIONAL/UNCHANGED because the qualitative ELF signature is suggestive and the claim may survive after a correct charge partition and a baseline comparison with ordinary metals, but the paper should not be accepted on the current quantitative evidence alone.","tokens_in":15734,"tokens_out":6749,"duration_ms":83530,"concrete_test":"On the self-consistent α-Mn density, run Bader analysis with only the 58 Mn nuclei as basins (no empty spheres) and a QTAIM search for non-nuclear maxima. If no non-nuclear attractor with basin charge >0.3 e coincides with the ELF=0.68 sites, the IAE charge is an artifact of pseudo-atom placement; if one does, the quantitative electride claim is supported. Separately, recompute Table 2 method (i) with cell-tiling Voronoi cells to verify charge conservation.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central question is not just whether ELF is high in voids, but whether the interstitial basins contain genuine anionic charge. Table 2 and Eq. (3) define the headline transfers. For α-Mn, 58 Mn atoms with Q_eff = +0.953 e and 16 IAEs with -1.645 e leave a unit-cell imbalance: 58×0.953 − 16×1.645 ≈ +29 e. Equivalently, the spherical regions account for 58×14.047 + 16×1.645 ≈ 841 of 870 valence electrons, leaving 29 e in unassigned space. Thus the abstract's -1.645 e value is not a well-defined charge transfer. The Bader values conserve charge, but the pseudo-atoms were placed exactly at the ELF maxima that they are supposed to validate; a Bader run without these empty spheres would assign those regions to Mn basins unless true non-nuclear maxima exist. No QTAIM critical-point analysis is reported. This circularity, together with the charge non-conservation, means the quantitative support for a native electride classification is not yet established. The 6–7% lattice errors are secondary; even with a perfect geometry, the charge analysis as presented cannot carry the claim.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports first-principles PBE+U calculations on three ambient-pressure phases of elemental manganese (α-, β-, and hexagonal Mn) and claims that all three are non-van der Waals electrides with spin-selective interstitial anionic electrons. The evidence presented includes spin-resolved ELF maps, charge-transfer values from spherical PAW-sphere integration and from Bader analysis, projected DOS/IDOS around the Fermi level, and AFM-versus-FM energy comparisons. The abstract's load-bearing quantitative claim is that each interstitial basin carries roughly -1.645e, -1.477e, and -1.083e in α-, β-, and hex-Mn, respectively.","tokens_in":15893,"tokens_out":5608,"duration_ms":63138,"significance":"If the claim is correct, classifying elemental manganese as a native magnetic electride under ambient conditions would be a notable result with implications for interstitial-electron chemistry, spintronics, and the search for electrides in transition metals. The paper has useful strengths: it considers three polymorphs, decomposes the ELF into spin channels, provides explicit AFM/FM energy differences consistent with experiment for α- and β-Mn, and attempts to cross-check charge transfers with two partitioning schemes and with IDOS. However, the central quantitative evidence for IAE charge is not currently reliable because the main spherical partition does not conserve electron number and because the IAE sites are defined by the same ELF maxima that are then used as evidence; the Bader and DOS analyses are performed with pseudo-atoms placed at exactly those positions. Until a charge-conserving, pseudo-atom-free topological analysis is provided, the electride classification rests on an inadequately supported descriptor.","major_comments":[{"comment":"The spherical-integration charge partition in Table 2 does not conserve electron number. For α-Mn, 58 Mn atoms with Q_eff = +0.953e account for +55.274e of charge loss, while 16 IAE spheres with -1.645e account for only 26.32e of charge gain; the integrated spherical charges sum to roughly 841e against 870 valence electrons, leaving about 29e in unassigned space. The same non-conservation, although smaller, occurs for β- and hex-Mn. Therefore the headline values of -1.645e, -1.477e, and -1.083e per IAE are not well-defined charge transfers; they depend on the chosen sphere radii and cannot by themselves carry the electride classification. Please replace or supplement this analysis with a space-filling, charge-conserving partition, or explicitly report the sphere radii and demonstrate that the missing charge does not affect the conclusion.","section":"Charge transfer analysis, Eq. (3), Table 2"},{"comment":"The identification of IAE sites is circular in its current form: the interstitial pseudo-atoms are placed at the maxima of the spin-resolved ELF, the same ELF maxima are then cited as evidence for non-nuclear anionic electrons, and the Bader and DOS analyses are performed with those pseudo-atoms present. A direct and necessary test is to perform a topological QTAIM analysis of the total charge density without inserting any empty spheres, to determine whether true non-nuclear (3,-3) critical points exist and whether their zero-flux basins contain substantial charge. Without this pseudo-atom-free analysis, the reported IAE Bader charges cannot be distinguished from an artifact of placing empty spheres at ELF-selected positions.","section":"Electron Localization Function; Charge transfer analysis; Ground state magnetic order"},{"comment":"The optimized AFM lattice parameter for α-Mn is 9.561 Å versus the experimental 8.91 Å, an overestimate of about 7.3%, and for β-Mn it is 6.670 Å versus 6.30 Å, an overestimate of about 5.9%, yet the text states these are in good agreement. Because the ELF, charge-transfer, and DOS results are all computed at these substantially expanded geometries, the quantitative electride descriptors may be affected. Please test the robustness of the central claim by repeating the key analyses at the experimental lattice constants or using a functional/U value that reproduces the experimental volumes, and show that the IAE topology and charges are stable.","section":"Crystal structures, Table 1"},{"comment":"The IDOS-derived populations are described as being in close agreement with the SC-VASP charges, but Table 3 gives IAE ΔN+ values of 1.558e, 1.422e, and 1.030e for α-, β-, and hex-Mn, whereas Table 2 lists Q_SC_tot of 1.645e, 1.477e, and 1.083e. The discrepancies are 5-9%, which is not close agreement, and both quantities are obtained from the same pseudo-atom spheres. Please reconcile the two sets of numbers or clarify the source of the discrepancy.","section":"Electronic Band structure and density of states, Table 3"}],"minor_comments":[{"comment":"The procedure for deriving the Wigner-Seitz radii from Bader volumes is not described; please provide the exact algorithm or a reference, and report all R_WS values for both Mn and IAE positions so that the spherical charge results are reproducible.","section":"Computational Methods, paragraph on Wigner-Seitz radii"},{"comment":"There are numerous typographical errors, including 'non-van der Waal' in the title, 'caluclated' in the caption of Fig. 2, and garbled axis labels in Fig. 5; the manuscript should be carefully proofread.","section":"Title and figures"},{"comment":"The ELF iso-surface threshold of 0.68 is described as a 'convenient maximum visualization threshold'; the number and shape of IAE regions can be threshold-dependent, so please show how the identified IAE positions and charges vary with the threshold or otherwise justify the chosen value.","section":"Electron Localization Function, Figs. 3 and 4"},{"comment":"The data availability statement says most data are in the main text and SI, but VASP is a commercial package; please deposit input files (POSCAR/CONTCAR, INCAR, KPOINTS) and analysis scripts (for Bader and IDOS) in an open repository so that the results can be reproduced.","section":"Data and Code Availability"}],"recommendation":"major_revision","confidential_remarks":"The paper's central quantitative claim rests on a non-conserving spherical charge partition and a circular ELF-to-pseudo-atom loop, and the AFM lattice parameters are 6-7% above experiment. These issues are load-bearing but appear fixable within the manuscript's scope: a pseudo-atom-free QTAIM analysis that conserves charge, together with a check of robustness to geometry, would either substantiate or refute the electride classification. I therefore recommend major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a plausible but not yet convincing claim that elemental Mn is an ambient-pressure non-vdW electride. The ELF pictures are striking, and the authors are the first to make this explicit claim, to my knowledge. But the quantitative charge evidence has a load-bearing flaw, and the geometry errors are large enough to worry about.\n\nWhat is actually new: spin-resolved ELF decomposition for α-, β-, and hex-Mn, with interstitial pseudo-atoms placed at ELF maxima, plus a comparison of two charge partitioning schemes. The paper is honest about the small magnetization at the interstitial sites. Also they compare AFM and FM states and benchmark Ueff against HSE06, which is reasonable practice.\n\nThe soft spots are serious. The SC-VASP spherical integration does not conserve charge: for α-Mn, 58 Mn atoms each lose 0.953 e but 16 IAEs gain only 1.645 e each. That leaves about 29 e unaccounted. So the -1.645 e per IAE from the abstract is not a well-defined charge transfer. The Bader numbers do conserve charge, but the basins were seeded at the ELF maxima, and there is no reported QTAIM critical-point analysis showing that the electron density actually has non-nuclear maxima at those points. Without that, the Bader \"IAE\" charges are not an independent validation. The lattice parameters for α- and β-Mn are 6-7% larger than experiment, which expands the interstitial voids and could artificially enhance electride character. And hex-Mn is not an ambient-pressure phase, so including it weakens the \"ambient pressure\" claim.\n\nThe \"spin-selective\" part is also overstated: the interstitial magnetic moments are below 0.2 μB, so the anionic electrons are essentially spin-compensated. The spin-resolved ELF just shows localization in both channels, which is not the same as spin-selectivity.\n\nNet: the idea is interesting and worth a serious referee, but the paper needs a pseudo-atom-free QTAIM/Bader analysis, a properly specified magnetic order, and an experimentally falsifiable prediction (work function, photoemission) before the electride classification can stand. I would send it out but with the expectation of major revision.\n\nRecommendation: engage with it if you work on electrides or Mn electronic structure; it will likely be cited as a provocative claim. For peer review, it deserves referee time, but it should not be accepted in its present form.","headline":"Plausible but not yet proven: the first explicit claim that elemental Mn is an ambient-pressure electride, undermined by a non-conserving charge partition and large lattice errors.","tokens_in":16560,"tokens_out":3398,"would_cite":false,"duration_ms":37450,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["71.20.-b","71.15.Mb"],"model":"deepseek-v4-flash","headline":"Elemental manganese, in each of its three ambient-pressure crystal phases, is claimed to be a non-van der Waals electride with spin-selective interstitial anionic electrons.","keywords":["electride","manganese","interstitial anionic electrons","electron localization function","spin-selective","density functional theory","Bader charge analysis","antiferromagnetism"],"falsifier":"If the electride claim is right, the spin-resolved ELF maxima and IAE basins must persist when the structure is re-optimized with a functional that reproduces the experimental lattice constants (for example HSE06, SCAN, or PBE with van der Waals corrections) and when the collinear antiferromagnetic order is replaced by a non-collinear or experimentally constrained spin texture. A calculation that removes the 6–7% lattice overestimate and finds no non-nuclear ELF maxima or IAE Bader charges above about $0.3e$ would falsify the central claim; alternatively, a high-resolution X-ray diffraction or Compton-profile experiment that sees no non-nuclear electron-density maxima at the predicted IAE sites would falsify it directly.","tokens_in":15406,"feed_emoji":"🧲","tokens_out":8846,"duration_ms":83832,"temperature":0.7,"pith_summary":"Elemental manganese is usually described as a transition metal bound by delocalized $d$-electrons. This paper claims that in all three ambient-pressure phases – cubic $\\alpha$, cubic $\\beta$, and hexagonal – a fraction of the electron density does not sit on the manganese atoms at all, but localizes in the empty interstitial pockets of the crystal, where it behaves as an anion. Using density-functional calculations with a Hubbard $U$, the authors find electron-localization-function maxima at non-nuclear sites, with Bader and sphere-integrated charges of about $-1.6$, $-1.5$, and $-1.1$ electrons per interstitial basin in $\\alpha$-, $\\beta$-, and hex-Mn. They further show these interstitial electrons contribute states at the Fermi level and are only weakly spin-polarized, while the antiferromagnetic order comes from the Mn $d$-electrons. If correct, this makes manganese a rare example of a pure element that is also an electride, potentially combining electride-like electron donation with magnetism.","feed_headline":"Manganese is an electride at ambient pressure","feed_subtitle":"Calculations find up to 1.6 electrons trapped in each crystal void, a native electride signature.","key_machinery":"The central object is the spin-resolved electron localization function (ELF), $\\mathrm{ELF}_\\uparrow$ and $\\mathrm{ELF}_\\downarrow$, computed from the Kohn–Sham kinetic-energy density. The ELF maps where same-spin electrons are locally scarce in kinetic energy, and its non-nuclear maxima define the interstitial anionic-electron (IAE) basins; these basins are then populated by pseudo-atoms (empty spheres), whose charge is quantified both by Bader zero-flux partitioning and by self-consistent spherical integration within adjusted Wigner–Seitz radii. The identity that carries the argument is that ELF maxima located away from both nuclei and Mn–Mn bonding regions, together with a negative effective charge on those basins and finite IAE density of states at the Fermi level, constitute the electride signature.","core_discovery":"The central claim is that elemental Mn under ambient pressure is a non-van der Waals electride, meaning an ionic solid in which the anions are cavity-trapped electrons rather than atomic species. The authors identify, in $\\alpha$-, $\\beta$-, and hex-Mn, zero-dimensional interstitial anionic-electron (IAE) sites at crystallographically defined voids, resolved separately in the spin-up and spin-down ELF channels. The effective charge transfer into each IAE basin is about $-1.645e$, $-1.477e$, and $-1.083e$ from two independent charge-partitioning schemes, and the IAE-projected density of states is finite at the Fermi level, so the interstitial electrons participate in the metallic low-energy electronic structure. The paper further distinguishes the electride electrons from the magnetism: the IAEs are almost spin-compensated (local moments below $0.2\\,\\mu_B$), whereas the antiferromagnetic ground state is carried by Mn local moments of about $\\pm 3.8$–$3.9\\,\\mu_B$. The conclusion is that Mn simultaneously is a native magnetic electride and an antiferromagnet, with the two behaviors arising from distinct electron populations.","pith_inferences":["If the same spin-resolved ELF analysis were applied to other elemental metals with large lattice voids, some may also qualify as native electrides, which would expand the electride family beyond compounds.","The paper does not report work functions, but known electrides have low work functions; measuring the work function of alpha-Mn would be a cheap indirect test of the claim.","The collinear antiferromagnetic order is not fully specified; recomputing the ELF and Bader charges with non-collinear magnetism or the full experimental spin structure would test the robustness of the IAE picture.","The negative charges of about one electron per void are large enough that a high-resolution X-ray diffraction measurement of the electron density, looking for non-nuclear maxima, could confirm or rule out the predicted interstitial anionic electrons."],"forward_implications":["A pure d-block metal can be a native electride, so electrides are not limited to compounds and intermetallics.","The interstitial electrons are electronically active: their states cross the Fermi level, so conduction and electride behavior coexist in the same material.","Because the IAE moments are an order of magnitude smaller than the Mn moments, the electride state and the antiferromagnetic order rest on different electron populations, opening a route to polarize the interstitial electrons without destroying the magnetic order.","The charge transfer scales inversely with the nearest-neighbor Mn–Mn bond length, linking the electride character to the geometric openness of the host lattice.","Each of the three phases has a distinct interstitial charge, giving a phase-dependent knob for electride strength in the same element."],"supporting_citations":[{"why":"Defines the electron localization function whose high, non-nuclear maxima identify the interstitial anionic-electron sites.","marker":"52"},{"why":"Supplies the grid-based Bader algorithm used as one of the two charge-partitioning schemes.","marker":"41"},{"why":"Provides the improved Bader allocation algorithm used for the charge-transfer numbers.","marker":"42"},{"why":"Gives the fast Bader decomposition implementation used in the charge analysis.","marker":"43"},{"why":"Provides the initial crystal structures of the three Mn phases used as starting geometries.","marker":"49"},{"why":"Provides the experimental antiferromagnetic structures and lattice parameters for alpha- and beta-Mn that anchor the comparison.","marker":"50"},{"why":"Justifies the PBE+U (Dudarev) treatment used for the localized Mn 3d electrons.","marker":"45"}],"fun_headline_variants":["Manganese is a spin-selective electride at ambient pressure","Elemental Mn is a non-van der Waals electride at 0 GPa","Interstitial electrons act as anions in manganese electride","Spin-selective electron localization reveals Mn as electride"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The computational model, whose lattice parameters for $\\alpha$- and $\\beta$-Mn are 6–7% larger than experiment, is treated as reliable enough for the electron-localization and charge-transfer values that define the electride classification.","fun_headline_variants_meta":{"raw":{"variants":["Manganese is a spin-selective electride at ambient pressure","Elemental Mn is a non-van der Waals electride at 0 GPa","Interstitial electrons act as anions in manganese electride","Spin-selective electron localization reveals Mn as electride"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000654,"raw_usage":{"total_tokens":3089,"prompt_tokens":1133,"completion_tokens":1956,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":749,"completion_tokens_details":{"reasoning_tokens":1883}},"tokens_in":749,"tokens_out":1956,"duration_ms":14681,"temperature":1.0,"reasoning_tokens":1883,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T00:46:38.330862+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"If the electride claim is right, the spin-resolved ELF maxima and IAE basins must persist when the structure is re-optimized with a functional that reproduces the experimental lattice constants (for example HSE06, SCAN, or PBE with van der Waals corrections) and when the collinear antiferromagnetic order is replaced by a non-collinear or experimentally constrained spin texture. A calculation that removes the 6–7% lattice overestimate and finds no non-nuclear ELF maxima or IAE Bader charges above about $0.3e$ would falsify the central claim; alternatively, a high-resolution X-ray diffraction or Compton-profile experiment that sees no non-nuclear electron-density maxima at the predicted IAE sites would falsify it directly.","supporting_citations":[{"cited_title":"Properties of Complex Inorganic Solids , pages=","cited_arxiv_id":null,"evidence_quote":"Supplies the grid-based Bader algorithm used as one of the two charge-partitioning schemes."},{"cited_title":"Journal of computational chemistry , volume=","cited_arxiv_id":null,"evidence_quote":"Provides the improved Bader allocation algorithm used for the charge-transfer numbers."},{"cited_title":"Journal of Molecular Structure: THEOCHEM , volume=","cited_arxiv_id":null,"evidence_quote":"Gives the fast Bader decomposition implementation used in the charge analysis."},{"cited_title":"Physical Review B , volume=","cited_arxiv_id":null,"evidence_quote":"Provides the initial crystal structures of the three Mn phases used as starting geometries."},{"cited_title":"Scientific reports , volume=","cited_arxiv_id":null,"evidence_quote":"Provides the experimental antiferromagnetic structures and lattice parameters for alpha- and beta-Mn that anchor the comparison."},{"cited_title":"Journal of computational chemistry , volume=","cited_arxiv_id":null,"evidence_quote":"Justifies the PBE+U (Dudarev) treatment used for the localized Mn 3d electrons."}],"review_version":1}