{"id":"70511e76-0ce4-4cf9-a8be-96f3d427a9c3","arxiv_id":"2608.04888","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Resonant coupling of interfacial states controls TMR in Cr2C/TMDC/Cr2C junctions, and 4% biaxial strain raises the TMR ratio to 540% for MoS2 and 496% for WS2.","lead":"This paper uses first-principles simulations to show that interface-induced electronic states and strain can boost the tunnel magnetoresistance (TMR) of all-two-dimensional magnetic tunnel junctions built from Cr2C electrodes and TMDC barriers. The results suggest practical routes for designing spin-selective tunneling devices.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The strained MoS2/WS2 TMR values rest on integrated conductances below 10^-6 e2/h with no k-grid or U-convergence checks, so the 540%/496% claims are not yet numerically grounded.","rationale":"The reader's weakest assumption — that U = 3.0 eV and the PBE functional, without convergence tests, adequately describe the interfacial states and tiny strained conductances — is exactly the load-bearing point. The paper is internally consistent and the mechanistic story (decay-rate-controlled hot spots plus interfacial-state resonance) is supported by the LDOS and conductance maps, which is real independent evidence. The thickness trends in Fig. 4 are coherent, and the stacking-energy analysis in Table S1 is complete enough to select the interfaces. However, the quantitative headline depends on G_AP values whose absolute scale is comparable to expected integration error on a 150×150 mesh, especially when conductance maps span eight orders of magnitude and hot spots are narrow. The two significant figures in Table S2 are insufficient to support TMR ratios quoted to three significant figures. These are addressable numerical robustness issues rather than a demonstrated flaw in the mechanism. Therefore the appropriate outcome is the reader's CONDITIONAL verdict; my read does not move that verdict, hence UNCHANGED. No ad hominem or theatrical language is intended; the concern is purely about numerical substantiation of a specific quantitative claim.","tokens_in":18369,"tokens_out":5619,"duration_ms":61403,"concrete_test":"Recompute the 3L MoS2 and WS2 junctions at 0%, 2%, 3%, and 4% strain with a 300×300 k∥ mesh and with U = 0 eV and U = 3 eV on Cr, keeping all other settings fixed, and report the spin-resolved G_AP values. If G_AP changes by more than 10%, or the TMR moves by more than 50 percentage points at any strain, the claimed 540%/496% values are not converged and the paper should either provide error estimates or soften the quantitative claim. Repeating with 100 K Fermi-Dirac smearing would additionally test broadening sensitivity.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative claim is the 4%-strain TMR ratio, whose denominator is the antiparallel conductance. For MoS2, Table S2 gives G_AP,↑ = 0.00027 and G_AP,↓ = 0.00025 in units of 10^-3 e2/h, i.e., G_AP ≈ 5.2×10^-7 e2/h, while G_P ≈ 3.2×10^-6 e2/h. The stated 540% is therefore a ratio of two very small Brillouin-zone integrals, each computed on a single 150×150 k∥ mesh with no convergence test and only two significant figures reported. At these magnitudes, a 10% error in G_AP changes the TMR by roughly 50 percentage points, and the rounding of G_AP,↑ versus G_AP,↓ alone introduces several percent uncertainty. No k-mesh-density, smearing-temperature, or Hubbard-U dependence is reported, and the spin-resolved conductance entering the denominator for WS2 (G_AP ≈ 2×10^-5 e2/h) is likewise small. The qualitative mechanism — resonant coupling of interfacial states, supported by the LDOS maps — can survive even if absolute numbers shift, but the headline 540% and 496% enhancement values cannot be validated without demonstrating that these tiny conductances are converged. The lack of deposited code or data makes independent verification impossible from the manuscript alone.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript presents first-principles NEGF calculations, using DFT+U (U = 3.0 eV on Cr) with PBE and SG15 pseudopotentials as implemented in QuantumATK, for Cr2C/MY2/Cr2C all-2D van der Waals magnetic tunnel junctions with MY2 = MoS2, WS2, MoSe2, and WSe2 and barrier thicknesses of 3, 5, 7, and 9 layers. It finds that the k-parallel-resolved conductance is suppressed near the Gamma point and enhanced at six off-Gamma hot spots, consistent with the barrier decay rates, and it attributes near-unity transmission channels in trilayer WS2, MoSe2, and WSe2 to resonant coupling of interfacial states that leave residual weight at the barrier center. The authors show that increasing barrier thickness attenuates these residual states, which monotonically reduces the TMR for MoS2 but produces a maximum at five layers for the other barriers. Applying biaxial tensile strain up to 4% to the trilayer junctions, the paper reports a strong increase in TMR (176% to 540% for MoS2, 98% to 496% for WS2) and a weaker response for MoSe2 and WSe2, which it attributes to spin-selective suppression of minority-spin interfacial states under strain.","tokens_in":18645,"tokens_out":7261,"duration_ms":84493,"significance":"As a qualitative mechanistic study, the paper has clear value: the k-parallel-resolved conductance maps in Fig. 2, the central-layer LDOS maps, and the decay-rate analysis in Fig. S3 form an internally consistent picture of interfacial-state-mediated tunneling, and the distinction between resonant channels and metallic spacer behavior (Sec. III A 2) is well argued. The TMR values are computed outputs of the NEGF calculation rather than fitted parameters, so there is no circularity in the central mechanism. If the quantitative strain claims are confirmed, the predicted 4%-strain TMR values would constitute useful design targets for MXene-based all-2D MTJs. However, the paper does not currently establish numerical convergence of the small conductances that determine the strain-enhanced TMR, and no code or data are deposited; hence the headline numbers are not yet robust. The mechanism itself is likely to survive even if the absolute values shift.","major_comments":[{"comment":"The headline strain-enhanced TMR ratios are built on conductances whose numerical reliability is not demonstrated. For MoS2 at 4% strain, Table S2 gives G_P,up = 0.0032, G_P,down = 0.000045, G_AP,up = 0.00027, and G_AP,down = 0.00025 in units of 10^-3 e2/h, so G_P = 3.2 x 10^-6 e2/h and G_AP = 5.2 x 10^-7 e2/h; the reported 540% ratio is extremely sensitive to the denominator, and a similar situation holds for WS2 (G_AP = 2 x 10^-5 e2/h) at 496%. The manuscript reports only a single 150 x 150 k-parallel mesh (Sec. II), with no convergence checks, no variation of the Fermi-Dirac smearing (300 K), and no dependence on the Hubbard U = 3.0 eV on Cr, which is the main empirical parameter in the Hamiltonian. I request k-mesh convergence tables (e.g., 150 x 150 versus 300 x 300), electronic-temperature dependence, and U-dependence (at least U = 0, 3, and 4 eV) for the 0% and 4% strained trilayer junctions, together with the raw conductance values with more significant figures.","section":"Sec. III C, Table S2, Sec. II"},{"comment":"The strain results assume that the unstrained ground-state interfacial stackings (M3 for MoS2 and MoSe2, M1 for WS2 and WSe2, Table S1) remain the most stable under biaxial tensile strain. Since the paper invokes the M1/M3 stacking difference to explain the barrier-dependent conductance (Sec. III A 2) and the mechanism of strain action is through modification of interfacial hybridization, the stacking energetics should be recomputed at the strain values plotted in Fig. 6 (at least at 4%). Without this check, the strain-enhanced TMR values are contingent on an unverified structural assumption.","section":"Sec. III C"}],"minor_comments":[{"comment":"The paper does not explicitly state the zero-bias condition for all TMR values; please state that Eq. (6) refers to zero bias and that finite-bias effects are outside the present scope.","section":"Sec. II, Eq. (6)"},{"comment":"The strained supercell parameters (in-plane lattice constant and the d_EB and d_BB values at 2-4% strain) are not reported; providing them in the Supplemental Material would aid reproducibility and interpretation of the LDOS changes.","section":"Sec. III C"},{"comment":"Adding distinct markers to each curve would improve readability, particularly for the 3L and 5L points where the curves for MoSe2 and WSe2 cross.","section":"Fig. 4"},{"comment":"Spin-orbit coupling is not discussed. For W-based TMDCs, SOC is sizable; a sentence justifying its neglect or providing an estimate of its effect would strengthen the spin-selective mechanism.","section":"Sec. II"}],"recommendation":"major_revision","confidential_remarks":"The stress-test criticism is the key issue: the strain-enhancement claim is not numerically grounded as it stands. I would not reject because the unstrained mechanism is coherent and the missing convergence tests are a standard, fixable request. I also note that the paper would benefit from depositing the raw conductance tables and LDOS maps, since the k-resolved maps are central to the argument."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The unstrained part of this paper is genuinely good. The authors give the first transport calculations for Cr2C/TMDC/Cr2C MTJs, and they build a coherent picture: the k-parallel conductance maps line up with the barrier LDOS, the near-unity transmission channels in WS2, MoSe2, and WSe2 correlate with residual states at the barrier center, and the thickness dependence follows naturally from the decay of those states. The decay-rate analysis explains the overall hot-spot pattern, and the LDOS-based mechanism is well supported by the figures. This is real work, and the mechanistic story for the trilayer unstrained junctions holds together.\n\nThe soft spot is the strained part. The 540% and 496% TMR ratios at 4% strain are ratios where the denominator (G_AP) is around 5e-7 e2/h for MoS2 and 2e-5 e2/h for WS2. These are tiny Brillouin-zone integrals computed on a single 150x150 k-mesh with no convergence test, and the values in Table S2 carry only two significant figures. A 10% error in G_AP changes the TMR by roughly 50 percentage points, and the rounding alone introduces several percent uncertainty. The stress-test note got this right: without k-grid density, smearing, or Hubbard-U dependence checks, the specific quantitative claims are not numerically grounded. The choice of U=3.0 eV is taken from prior Cr2C work and never varied, which matters for interfacial states that may be sensitive to the on-site Coulomb term.\n\nThat said, this is not a fatal flaw. The qualitative mechanism—strain suppresses minority-spin interfacial states, kills the resonant AP channels, and boosts TMR—does not depend on those exact numbers. The authors could reframe the strain section as a qualitative trend and the paper would still make a valid contribution. The lack of code or data deposition also weakens independent verification, but that alone is fixable.\n\nWho benefits? Anyone working on all-2D MTJs, MXene electrodes, or interface-state engineering. It deserves a serious referee, not a desk reject. The referee should ask for k-grid and U-convergence tests, error estimates on the lowest conductances, and either deposited data or a clear statement of why the numbers are stable. If those come back clean, the strain results become credible. If not, the qualitative mechanism still carries the paper.\n\nI'd bring it to a reading group and would cite it for the unstrained mechanism if I were working on TMDC-based MTJs, though I'd steer clear of quoting the 540%/496% figures until they're verified.","headline":"Solid mechanistic story for the unstrained Cr2C/TMDC junctions, but the headline strain-enhanced TMR numbers rest on conductances so small that the paper needs convergence evidence before I'd trust those values.","tokens_in":19214,"tokens_out":1977,"would_cite":true,"duration_ms":24480,"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 interface-induced metallic states, resonantly coupled across a TMDC barrier, control TMR in Cr2C-based all-2D junctions, and that 4% biaxial strain lifts TMR from 176% to 540% for MoS2 and from 98% to 496% for WS2.","keywords":["tunnel magnetoresistance","van der Waals magnetic tunnel junctions","interfacial states","Cr2C MXene","transition metal dichalcogenides","biaxial tensile strain","resonant tunneling","first-principles transport"],"falsifier":"Recompute the 4%-strained trilayer MoS$_2$ and WS$_2$ junctions with the on-site Coulomb repulsion on Cr set to 2.0 eV and 4.0 eV and with a denser $\\mathbf{k}_{\\parallel}$ grid, for example 300$\\times$300; if the TMR ratios do not remain near 540% and 496%, or if the minority-spin conductances shift by orders of magnitude, the strained-TMR claim is not numerically stable. An experiment that holds a Cr$_2$C/MoS$_2$/Cr$_2$C junction under 4% biaxial tension and sees no large TMR increase would also count against it.","tokens_in":18093,"feed_emoji":"🧲","tokens_out":8665,"duration_ms":90898,"temperature":0.7,"pith_summary":"This paper tries to establish that spin-dependent tunneling in all-two-dimensional magnetic tunnel junctions of the form Cr$_2$C/MY$_2$/Cr$_2$C is governed not only by the decay of evanescent waves in the barrier but also by metallic interfacial states that form at each electrode-barrier contact. The TMR ratio is the percentage resistance change when the two electrodes switch from parallel to antiparallel magnetization. For trilayer barriers of WS$_2$, MoSe$_2$, and WSe$_2$, the tails of these interfacial states reach the central layer and resonantly couple across the barrier, opening transmission channels near one conductance quantum at off-$\\Gamma$ hot spots. The paper further claims that 4% biaxial tensile strain selectively suppresses the minority-spin interfacial states in the sulfide junctions, raising TMR from 176% to 540% for MoS$_2$ and from 98% to 496% for WS$_2$, while MoSe$_2$ and WSe$_2$ respond only weakly. If correct, strain and barrier thickness become practical design knobs for all-2D spintronic devices.","feed_headline":"Strain lifts TMR from 176% to 540% in all-2D tunnel junctions","feed_subtitle":"Interface states set spin tunneling; 4% strain silences them to boost TMR nearly fivefold.","key_machinery":"The central object is the interfacial state: a metallic electronic state induced at the Cr$_2$C/TMDC interface by charge transfer and orbital hybridization, which decays into the otherwise insulating barrier. The load-bearing property is its residual weight at the barrier center, obtained from the $\\mathbf{k}_{\\parallel}$-resolved local density of states of the central TMDC layer; this weight measures whether the tails of the states from the two interfaces overlap. When the overlap is strong, the states resonantly couple across the barrier and create the near-unity transmission channels that set the conductance and TMR; when strain removes the minority-spin component of the interfacial states, it removes the antiparallel-channel bottleneck and raises TMR.","core_discovery":"The paper's central claim is that the tunnel magnetoresistance of all-2D Cr$_2$C/TMDC/Cr$_2$C junctions is governed by interface-induced metallic states rather than by bulk barrier decay alone. These interfacial states form by charge transfer and orbital hybridization at each Cr$_2$C-TMDC contact, penetrate into the barrier, and, in trilayer junctions of WS$_2$, MoSe$_2$, and WSe$_2$, leave a finite residual weight at the central layer. Where that residual weight is large, the left- and right-interface states resonantly couple and produce transmission channels near one conductance quantum ($e^2/h$) at six off-$\\Gamma$ hot spots; where it is small, as in MoS$_2$, those channels are suppressed. The authors then show that 4% biaxial tensile strain selectively removes the minority-spin interfacial states in the sulfide junctions, so the antiparallel conductance drops more than the parallel conductance and the TMR climbs from 176% to 540% for MoS$_2$ and from 98% to 496% for WS$_2$, while MoSe$_2$ and WSe$_2$ show only mild gains.","pith_inferences":["A cheap screening rule suggested by the paper: compute the Fermi-level LDOS of the central barrier layer alone; barriers whose central residual weight is large at the same $\\mathbf{k}_{\\parallel}$ points as the electrode states are likely to host the resonant channels that set TMR.","The same spin-selective strain mechanism might amplify TMR further in junctions whose electrodes have a cleaner spin polarization, where suppressing one spin channel at the interface would act on a more fully spin-polarized current.","Because the 4%-strain TMR values rest on conductances near $10^{-8}$ to $10^{-5}\\,e^2/h$, experimental confirmation may require thicker or alloyed barriers that keep the strain mechanism but bring the antiparallel conductance into a measurable range."],"forward_implications":["For trilayer MoS$_2$ junctions, applying 4% biaxial tensile strain raises the TMR ratio from 176% to 540%.","For trilayer WS$_2$ junctions the same strain raises TMR from 98% to 496%, with the minority-spin interfacial states almost fully suppressed.","Barrier thickness is a complementary tuning knob: five layers is optimal for WS$_2$, MoSe$_2$, and WSe$_2$, while MoS$_2$ loses TMR monotonically with thickness.","The residual weight of interfacial states at the barrier center acts as a transport fingerprint: conductance maps track the central-layer LDOS more closely than the interfacial LDOS.","High-transmission channels confined to a few $\\mathbf{k}_{\\parallel}$ points do not turn the barrier into a metallic spacer; tunneling character is preserved elsewhere in the Brillouin zone."],"supporting_citations":[{"why":"Documents strong interfacial coupling and orbital hybridization between a MXene and MoS2, the starting evidence that interfacial states form at such contacts.","marker":"[41]"},{"why":"Earlier Fe/MoS2/Fe study showing that interfacial states can dominate spin transport and set the thickness behavior this paper extends.","marker":"[50]"},{"why":"Shows that the k-resolved central-layer LDOS tracks the conductance, the correlation this paper uses to identify resonant coupling.","marker":"[54]"},{"why":"Provides the strain dependence of Cr2C-based monolayers, including the spin-dependent density-of-states shift that the strain mechanism relies on.","marker":"[45]"},{"why":"Earlier MXene Cr2C MTJ study; supplies the on-site Coulomb repulsion value U=3.0 eV used for Cr in this work.","marker":"[40]"},{"why":"Demonstrates that interface states can boost TMR when coupled across the barrier, the general mechanism here assigned to TMDC barriers.","marker":"[52]"}],"fun_headline_variants":["Strain lifts all-2D junction TMR to 540%","4% strain boosts TMR up to fivefold in 2D junctions","Interface states set TMR; strain resets them to 540%","All-2D junction strain triples TMR to 540%","Strain engineering lifts 2D TMR to 540%"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the density-functional calculation with the added on-site Coulomb repulsion U=3.0 eV on Cr captures the interfacial states and their penetration well enough that conductances as low as $4.5\\times10^{-8}\\,e^2/h$ and the resulting 540% TMR are real, not numerical noise.","fun_headline_variants_meta":{"raw":{"variants":["Strain lifts all-2D junction TMR to 540%","4% strain boosts TMR up to fivefold in 2D junctions","Interface states set TMR; strain resets them to 540%","All-2D junction strain triples TMR to 540%","Strain engineering lifts 2D TMR to 540%"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00075,"raw_usage":{"total_tokens":3450,"prompt_tokens":1163,"completion_tokens":2287,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":779,"completion_tokens_details":{"reasoning_tokens":2191}},"tokens_in":779,"tokens_out":2287,"duration_ms":22893,"temperature":1.0,"reasoning_tokens":2191,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T14:17:43.809299+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Recompute the 4%-strained trilayer MoS$_2$ and WS$_2$ junctions with the on-site Coulomb repulsion on Cr set to 2.0 eV and 4.0 eV and with a denser $\\mathbf{k}_{\\parallel}$ grid, for example 300$\\times$300; if the TMR ratios do not remain near 540% and 496%, or if the minority-spin conductances shift by orders of magnitude, the strained-TMR claim is not numerically stable. An experiment that holds a Cr$_2$C/MoS$_2$/Cr$_2$C junction under 4% biaxial tension and sees no large TMR increase would also count against it.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Documents strong interfacial coupling and orbital hybridization between a MXene and MoS2, the starting evidence that interfacial states form at such contacts."},{"cited_title":"Smidstrup, D","cited_arxiv_id":null,"evidence_quote":"Earlier Fe/MoS2/Fe study showing that interfacial states can dominate spin transport and set the thickness behavior this paper extends."},{"cited_title":"Masuda, H","cited_arxiv_id":null,"evidence_quote":"Shows that the k-resolved central-layer LDOS tracks the conductance, the correlation this paper uses to identify resonant coupling."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Earlier MXene Cr2C MTJ study; supplies the on-site Coulomb repulsion value U=3.0 eV used for Cr in this work."},{"cited_title":"Rungger, A","cited_arxiv_id":null,"evidence_quote":"Demonstrates that interface states can boost TMR when coupled across the barrier, the general mechanism here assigned to TMDC barriers."}],"review_version":1}