{"id":"cd019748-9026-48f6-b726-28c513021fac","arxiv_id":"2509.06823","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"DFT+NEGF calculations predict near-half-metallic perpendicular transport (spin polarization >90%) in Fe3GaTe2, Fe3GeTe2 and Fe4GeTe2, with Fe3GaTe2 closest to ideal half-metallicity.","lead":"This paper uses computer simulations to show that several atomically thin iron-germanium-telluride magnets, especially Fe3GaTe2, conduct almost only one spin direction when current flows perpendicular to their layers. This suggests they could be used as efficient electrodes in magnetic tunnel junctions for spintronics.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Half-metallic claim hinges on PBE placement of EF relative to the spin-down gap; a modest correlation shift would break the >90% SP numbers for F3GaT/F3GeT.","rationale":"I focused on the PBE Fermi-level placement because the central quantitative claim is a band-alignment statement: Eq. (3) converts the position of EF relative to the minority-spin transmission gap into the reported SP values, and the TMR numbers inherit this alignment. The paper's own discussion in Sec. IV C shows how sensitive the result is: F3GeT has EF only ~0.1 eV below the spin-down gap edge and still gets SP = 94%, while F3GaT is placed ~0.25 eV from the edge. PBE without U or SOC is not a controlled approximation for Fe d states; a self-energy correction of a few tenths of an eV is exactly the size that separates 'ideal half-metal' from 'highly polarized.' The paper tests robustness for F4GeT via Ref. 49 but not for the headline compound. This is a correctness risk, not merely a disagreement with consensus; a specific U or hybrid calculation would settle it. Secondary issues include the abstract's 'all compounds' claim, contradicted by the SP ≈ 2% for F5GeT(T) and by the absence of a transmission calculation for F5GeT(E), the conclusion's 'about 900%' versus the results' 'nearly 800%' TMR inconsistency, and a sentence at the end of Sec. IV C naming F3GeT rather than F3GaT as closest to ideal. These are presentation-level but should be corrected. The NEGF implementation and computational settings are otherwise standard, with good k-mesh convergence and a consistent methodology across compounds. I agree with the reader's weakest assumption and see no reason to change the CONDITIONAL verdict.","tokens_in":15473,"tokens_out":8065,"duration_ms":89658,"concrete_test":"Repeat the Smeagol transmission calculation for F3GaT and F3GeT with PBE+U (e.g., U_eff = 3 eV on Fe 3d) using the same geometries and the same 100×100 k-mesh, and recompute SP via Eq. (3). If EF leaves the spin-down transmission gap by more than ~0.1 eV, or if SP drops below 90%, then the near-ideal half-metal claim is not robust. Cross-check the energy of the F3GaT spin-down gap edges against the ARPES data of Ref. 72 to anchor the PBE band alignment.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The quantitative headline—SP = 97% for F3GaT, 94% for F3GeT, 92% for F4GeT—is a statement about the position of the Fermi level relative to the spin-down transmission gap, computed with PBE and without SOC or Hubbard U (Sec. III). This alignment is the load-bearing quantity: Eq. (3) converts it into SP, and the MTJ TMR inherits it. PBE is an approximate ground-state functional, while the transmission gap is an excited-state property, and in Fe d-electron systems self-energy corrections can shift EF or the minority d-band edge by several tenths of an eV. The paper itself shows the sensitivity: F3GeT has EF only ~0.1 eV below the spin-down conduction states (Sec. IV C), yet is assigned SP = 94%; F3GaT is deeper in the gap, with edges near -0.25 eV and +0.4 eV in Fig. 4(a), so its margin is ~0.25 eV. A correlation-induced shift of EF by 0.2–0.3 eV, or a self-energy correction moving the unoccupied minority band, would put F3GaT at the gap edge and downgrade it from 'ideal' to merely highly polarized. No U, hybrid, or SOC robustness test is reported for F3GaT; the only such discussion cites Ref. 49 for F4GeT. Thus the central claim is not yet protected against the main systematic error of the method.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports a systematic first-principles DFT+NEGF study of spin-dependent perpendicular transport in the van der Waals ferromagnets Fe3GaTe2, Fe3GeTe2, Fe4GeTe2, and Fe5GeTe2. It computes Fermi surfaces, k-resolved transmission coefficients, orbital-projected DOS, bulk spin polarization via Eq. (3), and TMR via Eq. (4) for model bilayer junctions with the vdW gap as barrier. The central claims are that F3GaT, F3GeT, and F4GeT have bulk spin polarizations of 97%, 94%, and 92%, respectively, that F3GaT is nearly an ideal half-metal with EF deep inside the spin-down transmission gap, and that model bilayer MTJs show TMR values of several hundred percent, with the F3GaT homobilayer reaching 800–900%.","tokens_in":15786,"tokens_out":6568,"duration_ms":74736,"significance":"If the quantitative claims are robust, this is a valuable consistent comparison of a promising family of vdW ferromagnets, and it strengthens the case for F3GaT as a high-Tc, high-anisotropy, near-half-metallic electrode material. The methodology has clear strengths: SP and TMR are direct outputs of DFT+NEGF with no parameters fitted to the target quantities; the same computational setup is applied across compounds; the Fermi-surface and transmission analyses are mutually consistent; and the F4GeT results agree with prior work. The PBE-level sensitivity of the band alignment is the main technical risk, and the manuscript contains internal inconsistencies that need correction before the claims can be accepted as stated.","major_comments":[{"comment":"For the experimentally derived F5GeT structure, the paper states that \"a full calculation of the transmission coefficient is not performed due to the large system size\" and that the Fermi surface \"strongly suggests a very high SP\". This is not a computed SP. The abstract and conclusion nevertheless claim that \"all compounds\" have SP exceeding 90%, while Fig. 3(a) contains no F5GeT value and the theoretical F5GeT structure gives only ~2%. The all-compounds claim is therefore unsupported for the most relevant experimental F5GeT structure. Either perform the transmission calculation with a coarser k-grid or explicitly restrict the claim to F3GaT, F3GeT, and F4GeT.","section":"Sec. IV.B, F5GeT paragraph"},{"comment":"The text states: \"In this regard, F3GeT already stands out as the material closest to ideal half-metallic behavior, with its EF lying deep within the spin-down gap.\" This directly contradicts the preceding analysis of Figs. 4(a)-(b), where F3GeT has EF only ~0.1 eV below the spin-down conduction states, whereas F3GaT has EF deep inside the spin-down gap (-0.25 eV to +0.4 eV), and it contradicts Fig. 3(a), which gives F3GaT the highest SP. If this is a typo, it must be corrected; as written it reverses the paper's ranking and weakens the central conclusion.","section":"Sec. IV.C, final summary"},{"comment":"The bulk SP values and the TMR values inherit their sensitivity from the PBE position of EF relative to the spin-down transmission gap. F3GeT has only ~0.1 eV of margin, and F3GaT has roughly 0.25 eV; these are the same order as typical self-energy or Hubbard-U corrections for Fe d-states. No robustness test (GGA+U, hybrid functional, SOC, or an explicit rigid shift) is reported for F3GaT or F3GeT; the only such test cited is Ref. [49] for F4GeT. Since \"near-ideal half-metallic\" is the headline claim, I request a sensitivity analysis, or at minimum a clear statement that the classification is PBE-level and could be altered by moderate corrections.","section":"Eq. (3) and Sec. IV.C"},{"comment":"The F3GaT homobilayer TMR is reported as \"nearly 800%\" in Sec. V but \"about 900%\" in the conclusion. This is a quantitative discrepancy in a headline result. Please determine the correct value and use it consistently throughout the text and Fig. 3(b).","section":"Sec. V vs. Sec. VI"}],"minor_comments":[{"comment":"Grammar: \"This findings underscore\" should be \"These findings underscore\".","section":"Abstract"},{"comment":"Typo: \"F3Ge3T\" should be \"F3GeTe2\"; also \"sizebale\" should be \"sizable\".","section":"Sec. VI"},{"comment":"The model MTJ uses Au s-orbital leads and the vdW gap as the barrier; this is a useful idealization but should be acknowledged more prominently in the abstract/conclusion so that the predicted TMR values are not read as predictions for realistic junctions with semiconducting barriers such as WS2 or h-BN.","section":"Sec. V"},{"comment":"The notation \"FnGe/GaT\" in the conclusion is ambiguous; consider writing Fe3GeTe2/Fe3GaTe2 or defining the family label explicitly.","section":"Sec. IV.A"}],"recommendation":"major_revision","confidential_remarks":"The main technical risk is the PBE Fermi-level alignment relative to the spin-down gap; the stress-test concern lands. The internal contradiction about F3GeT versus F3GaT in Sec. IV.C and the unsupported F5GeT claim in the abstract must be fixed before acceptance. I see no citation or novelty problems, and the paper's fit to the journal is appropriate."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is the first consistent DFT+NEGF comparison of F3GaT, F3GeT, F4GeT, and F5GeT for out-of-plane spin transport, and it does what previous literature didn't by putting all four on the same computational footing. It also corrects an earlier trend: F4GeT comes out slightly below F3GeT once computed with the same method, which is a useful clarification. The genuinely new result is F3GaT: SP ~97% with EF deep in the spin-down transmission gap, which is a strong electrode candidate given its 360 K Tc and large PMA. The Fermi-surface and transmission analyses are internally consistent, and the authors are honest about the F5GeT structural ambiguity.\n\nThe soft spots are real but not fatal. The headline numbers—97%, 94%, 92%—are PBE-level statements about where EF sits relative to the spin-down gap. The paper itself shows F3GeT's EF is only ~0.1 eV below the conduction states, so a modest correlation shift would knock it below 90%. F3GaT has more margin (~0.25 eV), but the authors never test it with U, a hybrid functional, or SOC, unlike F4GeT, where they cite Ref. 49. That is a gap in the central claim. Second, the abstract says 'all compounds' exceed 90%, but the experimental F5GeT structure is only inferred from its Fermi surface and the theoretical structure has SP ~2%; the conclusion correctly qualifies this, so the abstract should too. Third, the TMR is quoted as ~800% in the results and ~900% in the conclusion; those numbers need reconciling. Fourth, the bilayer MTJs use Au s-orbital leads, which is a deliberate model, not a quantitative prediction for a specific barrier.\n\nNone of this undermines the comparative ranking. The conclusion that F3GaT is the best of the four for perpendicular transport is well supported. The absolute SP and TMR values should be treated as PBE-level estimates until a correlation robustness check is done.\n\nWho this is for: anyone working on vdW MTJs or 2D ferromagnet electrodes. It deserves a serious referee. With the TMR inconsistency and abstract overreach fixed, and ideally a U or SOC check for F3GaT, it would be a solid publication.","headline":"A useful, consistent DFT+NEGF comparison that makes a credible case for F3GaT as a near-ideal half-metal; the quantitative SPs are PBE-level and the 'all compounds' framing overshoots.","tokens_in":16328,"tokens_out":2679,"would_cite":true,"duration_ms":29194,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["71.15.Mb","72.25.Ba","73.40.Gk","75.70.Cn"],"model":"deepseek-v4-flash","headline":"This paper claims that four layered van der Waals ferromagnets—and Fe3GaTe2 in particular—conduct perpendicular current nearly as half-metals, with spin polarizations above 90% and tunnel magnetoresistance of several hundred percent in mode","keywords":["van der Waals ferromagnets","Fe3GaTe2","Fe3GeTe2","half-metallic transport","spin polarization","magnetic tunnel junctions","tunnel magnetoresistance","NEGF transport"],"falsifier":"Measure the spin polarization of the out-of-plane conductance of Fe3GaTe2 at the Fermi level—for example by spin-resolved photoemission or point-contact Andreev reflection. A substantial minority-spin signal, or a DFT calculation with a more accurate treatment of Fe d electrons that moves the Fermi level out of the spin-down gap by more than about 0.2 eV, would falsify the near-ideal half-metallic claim.","tokens_in":15364,"feed_emoji":"🧲","tokens_out":9784,"duration_ms":106282,"temperature":0.7,"pith_summary":"This paper asks which of the van der Waals ferromagnets Fe3GeTe2, Fe4GeTe2, Fe5GeTe2, and Fe3GaTe2 is best suited as an electrode for magnetic tunnel junctions. Using first-principles density functional theory combined with non-equilibrium Green's function transport calculations, it shows that all of them, except Fe5GeTe2 in a particular theoretical structure, conduct electrons nearly as half-metals in the direction perpendicular to the layers. That means one spin channel carries current while the other has a transmission gap, giving bulk spin polarizations above 90%, with Fe3GaTe2 reaching about 97% and its Fermi level deep inside the minority-spin gap. The same property survives in model bilayer junctions where the van der Waals gap acts as a barrier, producing tunnel magnetoresistance of several hundred percent, near 800% for an Fe3GaTe2 homobilayer. The message is that Fe3GaTe2, which already has a Curie temperature above room temperature, is a realistic near-ideal half-metal for spintronics.","feed_headline":"Near-ideal half-metallic spin transport is found in Fe3GaTe2","feed_subtitle":"Bulk spin polarization tops 90% in four van der Waals magnets; model junctions reach TMR near 800 percent","key_machinery":"The working object is the spin-resolved transmission coefficient T_sigma(E_F, k_parallel), summed over the transverse Brillouin zone to give the linear-response conductance. What carries the argument is the spin-down transmission gap: in these layered compounds, out-of-plane current flows through Te 5pz states hybridized with Fe 3dz2 states, and the exchange splitting opens a gap in the minority channel. The paper shows that the position of the Fermi energy relative to this gap—not the size of the gap alone—determines the spin polarization and the tunnel magnetoresistance.","core_discovery":"The central claim is that perpendicular transport in this family is governed by a spin-down transmission gap rather than by a truly insulating minority band. In bulk Fe3GeTe2, Fe4GeTe2, and Fe3GaTe2 (and very likely in the experimentally reconstructed Fe5GeTe2), the spin-up transmission at the Fermi level is large while the spin-down transmission is essentially zero, yielding spin polarizations of about 94%, 92%, and 97% respectively. The gap comes from spin-split hybridization between Te 5pz and Fe 3dz2 orbitals perpendicular to the layers. The distinguishing feature of Fe3GaTe2 is that the Fermi energy lies deep inside this gap, whereas in Fe3GeTe2 it sits near the gap edge and in Fe4GeTe2","pith_inferences":["Since replacing Ge with Ga in F3GeT is what moves the Fermi level deep into the gap, alloying or electrostatic gating across the FnGeTe2 family may be a tunable knob for optimizing spin polarization.","A direct experimental check would be spin-resolved photoemission or point-contact Andreev reflection on exfoliated Fe3GaTe2; finding minority-spin weight at the Fermi level would downgrade the near-ideal claim.","The model junctions treat the bare van der Waals gap as the tunnel barrier and use simplified s-orbital leads; real insulating barriers such as WS2, WSe2, or h-BN may alter interfacial transmission, so the predicted TMR should be tested with explicit barrier calculations.","If confirmed, the same gap mechanism might be sought in isostructural doped variants, for example Fe3Ga1-xGexTe2, to map where half-metallicity disappears."],"forward_implications":["Fe3GaTe2 becomes the standout candidate: near-ideal half-metallic conductance (SP around 97%), a Curie temperature above room temperature, and strong out-of-plane anisotropy make it a plausible all-van-der-Waals MTJ electrode.","Homobilayer vdW-gap junctions of Fe3GaTe2 are predicted to show TMR of roughly 800–900%, several times larger than the reported 213–340% in existing F3GaT/WS2 and F3GaT/WSe2 devices.","Heterojunctions made from two different FnGe/GaTe compounds keep TMR around 400–500%, suggesting independently switchable layers are practical.","The near half-metallicity is shared by F3GeT and F4GeT, so the family is robust across stoichiometry, but the theoretical P3m1 F5GeT structure is a counterexample with SP about 2%; half-metallicity depends on the experimentally derived UUD structure.","A consistent computational setup changes the ordering from earlier comparisons: F3GeT is not lower than F4GeT in spin polarization when both are treated identically."],"supporting_citations":[{"why":"Supplies the experimental crystal structure, lattice parameters, and above-room-temperature Curie temperature of Fe3GaTe2, the compound identified as near-ideal.","marker":"[33]"},{"why":"Supplies the experimental structures and lattice parameters of the FenGeTe2 family (Fe3GeTe2, Fe4GeTe2, Fe5GeTe2).","marker":"[66]"},{"why":"Earlier calculation predicting high spin polarization in Fe3GeTe2 that motivates the consistent comparison made here.","marker":"[28]"},{"why":"Authors' prior DFT+NEGF study of Fe4GeTe2, which suggested near half-metallic transport and is extended to the family.","marker":"[49]"},{"why":"Experimental source for the reconstructed Fe5GeTe2 structure with UUD ordering used in the high-polarization calculation.","marker":"[31]"},{"why":"Theoretical work showing the UUD configuration is consistent with STM, justifying the choice of the experimental Fe5GeTe2 structure.","marker":"[43]"},{"why":"Prior transport study that assumed the P3m1 Fe5GeTe2 structure; its low spin polarization contrasts with the experimental structure.","marker":"[44]"},{"why":"Experimental F3GaT/WS2/F3GaT magnetic tunnel junction giving TMR of 213% and spin polarization 72% at 10 K, the benchmark for the predicted TMR.","marker":"[35]"},{"why":"Experimental F3GaT/WSe2/F3GaT junction reporting TMR up to 340% at 10 K and 50% at 300 K, another benchmark for the predicted values.","marker":"[36]"},{"why":"Supplies the implementation of the DFT+NEGF method used to compute all transmission coefficients and spin polarizations.","marker":"[52]"}],"fun_headline_variants":["Fe3GaTe2 shows ideal half-metallic transport","Spin polarization exceeds 90% in vdW ferromagnets","TMR near 800% in magnetic junctions from vdW magnets","Half-metallic behavior in Fe3GaTe2 for spintronics","Van der Waals ferromagnets achieve high spin polarization"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The conclusions rest on the calculated Fermi-level position in relation to the spin-down transmission gap; the approximate exchange-correlation functional used for Fe d electrons is not exact, and a shift of the Fermi level by a few tenths of an eV would reduce some of these compounds from half-metallic to merely highly spin-polarized.","fun_headline_variants_meta":{"raw":{"variants":["Fe3GaTe2 shows ideal half-metallic transport","Spin polarization exceeds 90% in vdW ferromagnets","TMR near 800% in magnetic junctions from vdW magnets","Half-metallic behavior in Fe3GaTe2 for spintronics","Van der Waals ferromagnets achieve high spin polarization"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000351,"raw_usage":{"total_tokens":1778,"prompt_tokens":799,"completion_tokens":979,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":543,"completion_tokens_details":{"reasoning_tokens":888}},"tokens_in":543,"tokens_out":979,"duration_ms":11045,"temperature":1.0,"reasoning_tokens":888,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-04T23:00:25.766668+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the spin polarization of the out-of-plane conductance of Fe3GaTe2 at the Fermi level—for example by spin-resolved photoemission or point-contact Andreev reflection. A substantial minority-spin signal, or a DFT calculation with a more accurate treatment of Fe d electrons that moves the Fermi level out of the spin-down gap by more than about 0.2 eV, would falsify the near-ideal half-metallic claim.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the experimental structures and lattice parameters of the FenGeTe2 family (Fe3GeTe2, Fe4GeTe2, Fe5GeTe2)."},{"cited_title":"Halder, D","cited_arxiv_id":null,"evidence_quote":"Authors' prior DFT+NEGF study of Fe4GeTe2, which suggested near half-metallic transport and is extended to the family."},{"cited_title":"Ershadrad, S","cited_arxiv_id":null,"evidence_quote":"Theoretical work showing the UUD configuration is consistent with STM, justifying the choice of the experimental Fe5GeTe2 structure."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Prior transport study that assumed the P3m1 Fe5GeTe2 structure; its low spin polarization contrasts with the experimental structure."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Experimental F3GaT/WS2/F3GaT magnetic tunnel junction giving TMR of 213% and spin polarization 72% at 10 K, the benchmark for the predicted TMR."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Experimental F3GaT/WSe2/F3GaT junction reporting TMR up to 340% at 10 K and 50% at 300 K, another benchmark for the predicted values."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the implementation of the DFT+NEGF method used to compute all transmission coefficients and spin polarizations."}],"review_version":1}