{"id":"7e753043-9646-41a8-bd5a-6fcce43ca878","arxiv_id":"2607.15728","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Interface termination and h-BN/graphene insertion are predicted to select the tunneling spin channel in CrSb/In2Se3 altermagnetic junctions, inverting the P/AP resistance hierarchy and yielding 4-8 nonvolatile resistance states with giant computed TMR/TER.","lead":"Computer simulations predict that the atomic termination at the interface between altermagnetic CrSb and ferroelectric In2Se3 controls which spin channel tunnels, flipping the low/high-resistance correspondence and enabling up to eight nonvolatile resistance states with giant computed TMR (~1600-38000%) and TER (~2200-418000%). It matters because it proposes interface symmetry, not just the bulk magnet, as a design knob for stray-field-free multistate spintronic memory.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Claimed termination-reversal rule is internally contradicted: the headline Cr-S/2·h-BN/In2Se3 device shows P as low-resistance, opposite to Fig. 2(b)'s 'symmetric → P high'.","rationale":"Good faith: the paper contains a coherent computational study, sensible methods, and a physically motivated mechanism. The issue I isolate is not external robustness (strain, SOC, stacking, gating) but an internal inconsistency in the central claim. The reader's weakest_assumption emphasizes idealized-interface realism; I agree those are secondary risks. The reader's rationale point (1) already notes some contradictions, so my read partially agrees. I sharpen the finding: the single worst case is the device used for the headline 1626% TMR, where Fig. 2(b)'s rule fails in both polarizations. If the rule has this many exceptions in the paper's own data, the abstract's universal statement is not supported. A revised version could restrict the rule to certain barrier/insertion combinations or identify the actual controlling variable (e.g., the local interfacial Cr-moment alignment rather than S/A label). Since that is an addressable revision, I retain the reader's CONDITIONAL verdict rather than moving to REJECT; the concrete sign-table check would determine whether a more severe verdict is needed.","tokens_in":17304,"tokens_out":8456,"duration_ms":70082,"concrete_test":"Compile a full sign table from the printed data: for each row of Tables 1–2 and S1–S4, compute sgn(T_P − T_AP) and compare with Fig. 2(b)'s prediction for that termination class and polarization. Also list explicitly for the headline Cr-S/2·h-BN/In2Se3 junction whether P or AP is low-resistance. No new DFT is needed; this uses already-reported T_tot values. If, as the main-text numbers suggest, more than a handful of rows violate the rule, the paper must replace the universal termination-reversal statement with a configuration-dependent rule or a quantitative criterion based on calculated interfacial Cr-moment alignment.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Tables 1 and 2 list raw spin-resolved transmissions, so the P/AP resistance hierarchy can be read off directly. Under the paper's own definition (Eq. 5, TMR = |T_P−T_AP|/min{...}), the absolute value hides the sign; the raw T_tot columns do not. Take the highest-TMR monolayer device, Cr-S/2·h-BN/In2Se3 (Table 2 row vi): for P→, T_P=4.97×10⁻⁵ and T_AP=8.58×10⁻⁴ — P is the low-resistance state, ~17× more conductive than AP; for P←, T_P=2.23×10⁻⁴ and T_AP=8.06×10⁻⁴, again P low. This is exactly opposite to Fig. 2(b)'s universal rule that symmetric terminations make P high-resistance. The contradiction is not confined to inserted layers: direct Cr-S at PFE← has T_P=1.02×10⁻¹ vs T_AP=6.89×10⁻² (P low), while same interface at PFE→ has P high. Asymmetric Sb-A and Cr-Sb-A are P-high in both polarizations, violating the 'asymmetric → P low' branch. Thus termination symmetry is not the selector of the P/AP hierarchy; the hierarchy depends jointly on termination, insertion layer, and polarization. Since the abstract's central claim is precisely that symmetric/asymmetric terminations reverse the correspondence, this internal inconsistency is the load-bearing weakness.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports first-principles NEGF-DFT calculations of CrSb/α-In2Se3 altermagnetic multiferroic tunnel junctions, with h-BN/graphene insertion layers, claiming that the symmetry of the interface terminations (symmetric vs. asymmetric) reverses the TMR resistance hierarchy between parallel and antiparallel Néel-vector configurations. The authors report giant TMR (up to 1626% at equilibrium, 9576% with Fermi-level shifting) and TER (up to 2206%, 4144%) for monolayer barriers, and up to eight resistance states using bilayer In2Se3 with TMR and TER reaching 3.77×10^4% and 4.18×10^5%. The underlying microscopic mechanism is argued to be interface-selected spin-channel matching controlled by actual interfacial Cr-moment alignment, not the bulk altermagnetic splitting alone.","tokens_in":17573,"tokens_out":6045,"duration_ms":49166,"significance":"If correct, the proposed interface-symmetry-controlled spin-channel matching would establish a new design principle for altermagnetic multiferroic tunnel junctions and expand the multistate memory concept. The work uses standard, carefully parameterized DFT and NEGF calculations (500 eV cutoff, 13×13×1 and 100×100×1 k-meshes, DZP basis), and the transmission data are provided in raw form, which is commendable for reproducibility. However, the central claim of a universal termination-induced inversion of the P/AP resistance hierarchy is directly contradicted by the paper's own Tables 1 and 2, as detailed below. Because this rule is the main conceptual advance advertised in the abstract and Fig. 2(b), the paper's core conclusion is not supported.","major_comments":[{"comment":"The claimed universal rule—symmetric terminations make the P configuration high-resistance and AP low-resistance, with the reverse for asymmetric terminations—is internally contradicted by the reported transmissions. For the symmetric Cr-S interface in Table 1, PFE← gives T_P=1.02×10^-1 > T_AP=6.89×10^-2, i.e., P is the low-resistance state, the opposite of the rule. Similarly, Table 1 Sb-S with PFE← has T_P=4.58×10^-2 > T_AP=4.37×10^-2. For asymmetric Sb-A, PFE← gives T_P=3.05×10^-2 < T_AP=4.00×10^-2 (P high), also opposite to the expected reversed hierarchy. Table 2, which all use the symmetric Cr-S interface, shows the rule only intermittently: Gr/In2Se3 gives P low for both polarizations (T_P=1.24×10^-2 vs T_AP=4.74×10^-3 for PFE→), and 2·Gr/In2Se3 likewise (T_P=3.26×10^-3 vs T_AP=2.12×10^-3). These examples are not isolated; the hierarchy is not controlled by termination symmetry al","section":"Fig. 2(b) vs Tables 1 and 2"},{"comment":"The TMR definition uses an absolute value, which discards the sign of the resistance hierarchy. The raw T_tot columns in Tables 1 and 2 are therefore essential, and they reveal the above contradictions. The authors should report the actual high/low-resistance assignment (e.g., whether P or AP has higher conductance) for each state, rather than only the absolute TMR. Without this, the paper's interpretation cannot be checked from the TMR numbers alone.","section":"Eq. (5) and raw transmission columns"},{"comment":"The claim that TMR reaches 9576% at E−E_F = −0.37 eV and TER 4144% at −0.06 eV is presented as 'Fermi-level shifting' enabling even larger ratios. However, no physical mechanism for shifting the Fermi level in a metallic CrSb electrode is given—electrostatic gating of a metal is not straightforward, and no gated calculation is performed. These off-E_F values are therefore not device predictions but simply evaluations of an energy-dependent transmission function. If these numbers are to be advertised in the abstract, a concrete gating/doping model is required.","section":"Fig. 5 and Fermi-level shifting"},{"comment":"The giant TMR/TER values rely on transmission contrasts of several orders of magnitude (e.g., Table 2 2·h-BN/In2Se3: T~10^-5 vs 10^-4). These predictions assume idealized coherent interfaces: the h-BN and graphene supercells are strained by ~4–6% to match CrSb, and the energetically preferred stackings from Fig. S2 are taken as the device geometry. No account is made of possible surface reconstruction, intermixing, or relaxation of the CrSb surface in contact with the insertion layers. Since the proposed microscopic mechanism operates at the interface, the quantitative predictions—and even the qualitative hierarchy—could change under more realistic interfacial structures. This is a limitation that should be stated and ideally tested with at least one alternative stacking or a short molecular-dynamics relaxation.","section":"Structural idealization"}],"minor_comments":[{"comment":"The paper contains several typographical and formatting issues, including inconsistent use of 'Néel' vs 'N√©el' in the extracted text, and the figure captions are sometimes terse. Please proofread carefully.","section":"General"},{"comment":"The schematic in Fig. 2(b) is difficult to interpret, especially the petal-like symbols and the meaning of 'favorable spin-channel matching'. A clearer graphical definition of the resistance hierarchy (and its dependence on termination) would help readers.","section":"Fig. 2(b)"},{"comment":"The tables list TMR and TER to one decimal place, but some entries (e.g., TER 0.3%, 0.2%) appear marginally different from rounding. It would be useful to give the raw transmission values with more significant figures or at least the conductance ratio explicitly.","section":"Table 1 and Table 2"},{"comment":"The choice of U=0 for CrSb is justified by comparison to ARPES, but the effect of possible U on the interfacial transmission is not discussed. A sentence on the sensitivity of transport coefficients to U would be informative.","section":"Methods"},{"comment":"Several references are self-citations of the authors' previous work (e.g., Refs. 5, 6, 11, 40, 42, 43). While these are relevant, the paper would benefit from citing more independent experimental and theoretical studies of CrSb-based tunnel junctions beyond the current group.","section":"References"}],"recommendation":"reject","confidential_remarks":"The central claim of the paper—that symmetric and asymmetric terminations universally reverse the P/AP resistance hierarchy—is contradicted by the paper's own data in Tables 1 and 2. This is not a minor presentation issue; it is the core result advertised in the abstract and the basis for the proposed design principle. The calculations appear carefully performed, but the interpretation is not supported. Even if the authors were to restrict the claim to specific barrier configurations, the current manuscript would require a substantial rewrite and a re-evaluation of the significance. I recommend rejection rather than major revision because the load-bearing message is incorrect as stated."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First, what's actually new: this is the first systematic scan I know of that treats termination (Cr vs Sb, symmetric vs asymmetric) and h-BN/graphene insertion layers as coupled knobs for spin-channel matching in CrSb altermagnetic MFTJs. The PDOS and k∥-resolved transmission maps give a coherent picture of why certain interfaces favor one spin channel, and the monolayer–bilayer comparison with In2Se3 polarization configurations is a reasonable route to four- and eight-state devices. The numerics look careful: 500 eV cutoff, 13×13×1 and 100×100×1 k-meshes, DZP basis, and U=0 justified by ARPES comparison.\n\nThe load-bearing claim, however, is that symmetric terminations make P high-resistance and asymmetric terminations invert that — stated as general in the abstract and Fig. 2(b). The paper's own Table 1 contradicts it. For the direct Cr-S junction with PFE←, T_P = 1.02×10⁻¹ and T_AP = 6.89×10⁻², so P is the low-resistance state. The rule only holds for some configurations, e.g., the 2·h-BN/In2Se3 row where T_P = 4.97×10⁻⁵ and T_AP = 8.58×10⁻⁴, P high. So the hierarchy depends jointly on termination, insertion layer, and polarization. The universal statement is an overclaim.\n\nOther soft spots are smaller. The |Δ|/min normalization in Eqs. 5 and 7 inflates the ratios and hides whether P or AP is more conductive; conventional TMR denominators would make the numbers more modest and also expose the sign. The off-Fermi-level maxima (“Fermi-level shifting”) are just transmission peaks, with no gating model for a metallic electrode. No SOC at the heavy-Sb interface is a real concern, though collinear is a defensible first pass given the exchange-driven splitting. The eight-state bilayer claim is “in principle” in the text but asserted in the abstract, and the stability of the four bilayer polarization configurations is not demonstrated. No code or data are shipped.\n\nThese are addressable. The qualitative idea — that interface termination and insertion layers select spin channels and can invert the P/AP hierarchy — could survive, but it needs to be stated with its regime of validity, not as a universal rule. For a spintronics or materials theory audience, this is a useful catalog of interface effects in CrSb-based MFTJs and a good example of how parameter scans can overstate a rule. I would send it to peer review, but with major revision: narrow the claim, recompute ratios under conventional definitions, add sensitivity checks (SOC, strain, stacking, U), and show the full 8-state resistance matrix.","headline":"A well-converged computational parameter sweep of CrSb/In2Se3 multiferroic tunnel junctions whose headline 'termination reverses P/AP hierarchy' rule is contradicted by the paper's own transmission tables — needs revision, but deserves peer review.","tokens_in":18259,"tokens_out":3530,"would_cite":false,"duration_ms":29870,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper claims that the termination symmetry of CrSb/α-In2Se3 interfaces, not the bulk altermagnetic splitting alone, selects the tunneling spin channels, reverses the parallel/antiparallel resistance hierarchy, and yields four or eight n","keywords":["altermagnetism","CrSb","multiferroic tunnel junctions","tunneling magnetoresistance","tunneling electroresistance","spin-channel matching","interface termination","alpha-In2Se3"],"falsifier":"A noncollinear or spin-orbit-inclusive calculation of the same junctions—or an epitaxial sample with controlled termination—that fails to show the termination-dependent reversal of the P/AP resistance hierarchy would refute the central claim.","tokens_in":17021,"feed_emoji":"🧲","tokens_out":6947,"duration_ms":55308,"temperature":0.7,"pith_summary":"The paper aims to establish that, in CrSb-based altermagnetic multiferroic tunnel junctions, the dominant tunneling channels are selected by the atomic symmetry of the electrode/barrier interface rather than by the bulk altermagnetic spin splitting alone. It shows that Cr versus Sb termination and the insertion of h-BN or graphene layers determine spin-channel matching, while ferroelectric polarization of the In2Se3 barrier reshapes the electrostatic barrier. As a result, symmetric and asymmetric interfaces reverse which Néel-vector configuration (parallel or antiparallel) gives high versus low resistance. Monolayer barriers yield four nonvolatile resistance states with TMR up to 1626% (9576% with Fermi-level shifting) and TER up to 2206% (4144%), while bilayer barriers give eight states with TMR up to 3.77×10^4% and TER up to 4.18×10^5%. If true, interface symmetry becomes a practical design knob for stray-field-free, multistate spintronic memory.","feed_headline":"Termination flips which CrSb spin state is low-resistance","feed_subtitle":"Interface termination decides which spin channel tunnels, unlocking 4 or 8 states.","key_machinery":"The carrying mechanism is spin-channel matching between the two electrode/barrier interfaces: Cr/Sb termination and h-BN or graphene insertion layers set which spin channel has continuous spectral weight through the tunneling barrier, while ferroelectric α-In2Se3 polarization reshapes the electrostatic barrier and redistributes the k‖-resolved transmission. The calculations combine density functional theory with nonequilibrium Green's function transport, evaluating spin-resolved transmission and current through Landauer expressions; the termination-dependent reversal of the P/AP resistance hierarchy is the central discriminating prediction.","core_discovery":"The central claim is that the resistance state is fixed by the actual alignment of the interfacial Cr moments and the spin-channel matching it permits, not by the nominal parallel/antiparallel label of the electrode Néel vectors. Symmetric interfaces make the P configuration high-resistance and AP low-resistance; asymmetric interfaces reverse this, for both ferroelectric polarizations. Spin-resolved transmission maps locate conductance hot spots in a single spin channel near the Fermi level, and ferroelectric polarization redistributes these hot spots, producing both tunneling magnetoresistance and tunneling electroresistance. Extending the barrier to bilayer In2Se3 introduces up/down/head-t","pith_inferences":["The hierarchy-reversal rule likely extends to other altermagnet/barrier combinations, but the sign and magnitude will be set by which k‖ regions a specific interface couples, so it must be recomputed for each material pair.","The off-Fermi-level TMR/TER boosts assume a controllable shift of the transport window in a metallic electrode; practical routes such as doping or bias protocols are not modeled in the paper.","A direct experimental check is to grow junctions with symmetric and asymmetric terminations and compare the sign of TMR; observing the predicted reversal would confirm spin-channel matching as the controlling variable.","The computed four-to-five-order-of-magnitude transmission contrasts depend on atomically perfect interfaces; roughness, reconstruction, or intermixing will shrink the ratios before altering the qualitative hierarchy."],"forward_implications":["Cr-terminated CrSb interfaces support stronger spin-selective tunneling than Sb-terminated ones, making termination a first-order performance knob.","Switching between symmetric and asymmetric terminations toggles the sign of TMR, adding a structural degree of freedom for nonvolatile memory without changing the magnetic order.","Monolayer α-In2Se3 barriers give four resistance states, with maximum TMR 1626% and TER 2206%, and shifting the transport window can raise these to 9576% and 4144%.","Bilayer In2Se3 barriers give eight resistance states, with maximum TMR 3.77×10^4% and TER 4.18×10^5%.","Finite-bias transport retains robust spin filtering and spin-polarized currents, so the multistate response is not only an equilibrium feature."],"fun_headline_variants":["Termination dictates which CrSb state is low-resistance","Altermagnetic junction: interface termination flips resistance","Spin-channel matching unlocks 4 or 8 CrSb resistance states","Ferroelectric + termination: up to 8 resistance states","CrSb/In2Se3: termination controls spin-channel tunneling"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The numerical predictions assume the idealized, perfectly ordered epitaxial interfaces used in the calculations are the actual device structures; if the real CrSb surface relaxes, reconstructs, or intermixes with In2Se3, the enormous resistance ratios degrade first, even if the qualitative termination dependence survives.","fun_headline_variants_meta":{"raw":{"variants":["Termination dictates which CrSb state is low-resistance","Altermagnetic junction: interface termination flips resistance","Spin-channel matching unlocks 4 or 8 CrSb resistance states","Ferroelectric + termination: up to 8 resistance states","CrSb/In2Se3: termination controls spin-channel tunneling"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000365,"raw_usage":{"total_tokens":1855,"prompt_tokens":855,"completion_tokens":1000,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":599,"completion_tokens_details":{"reasoning_tokens":915}},"tokens_in":599,"tokens_out":1000,"duration_ms":9289,"temperature":1.0,"reasoning_tokens":915,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T22:30:30.445237+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A noncollinear or spin-orbit-inclusive calculation of the same junctions—or an epitaxial sample with controlled termination—that fails to show the termination-dependent reversal of the P/AP resistance hierarchy would refute the central claim.","supporting_citations":[],"review_version":1}