{"id":"f953f9e3-7c34-42e7-bc12-d80a51e16bbf","arxiv_id":"2412.17949","paper_version":4,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Topological junction states in graphene nanoribbons bind NO2 more than twice as strongly as the rest of the ribbon and produce a measurable current change, suggesting a new sensing mechanism.","lead":"This paper uses quantum simulations to show that topological junction states in graphene nanoribbons can act as strong binding sites for gas molecules. It proposes a new type of gas sensor based on these states, with NO2 as a test case.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The central topological-chemistry claim is not yet separated from ordinary junction chemistry because no Z2=0 control junction is computed and the surviving state in the methyl-terminated device is asserted rather than verified.","rationale":"The reader's verdict identifies the survival of the TJS after methyl termination as the weakest assumption, and that is a real concern. My stress-test refines this into a sharper issue: the causal link between the Z2 invariant and the observed adsorption/transport enhancement is never directly demonstrated, because no topologically trivial control junction is computed. Even if the TJS survives termination, the enhanced reactivity could be an ordinary property of the A60/AGNR interface, not of the topological bulk invariant. The paper deserves credit for a systematic TB classification, reproducible code and data, and internally consistent DFT/transport results. But the central 'topological chemistry' claim requires the control calculation to rule out non-topological junction chemistry. Since the paper already has the CONDITIONAL verdict, my concern does not move the verdict; it strengthens the justification for the outstanding condition.","tokens_in":13431,"tokens_out":11230,"duration_ms":117439,"concrete_test":"Repeat the adsorption and transport calculations on a control double junction built from a topologically trivial member of the same A60 family, e.g. A60(2,2,5,5)-AGNR(7), keeping the same fragment lengths, methyl termination and alignment point as the Z2=1 A60(2,2,4,4)-AGNR(7) device. If the trivial control shows comparable adsorption energy and current enhancement, the topological character is not the cause; if it shows no enhancement while the Z2=1 junction does, the topological attribution is supported. In parallel, compute the intercellular Zak phase/Z2 invariant for a periodic supercell that includes the methyl-terminated A60/AGNR unit cell to verify that the TJS survives saturation.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central claim is that the localized topological junction state (TJS) itself, not just any interface state, boosts NO2 adsorption and doubles the transport current. All chemistry and transport results are computed on the methyl-terminated double junction (DJ) of Fig. 2, whose Z2 invariant is never recomputed; the text states only 'we expect one TJS per junction similar to Figure 1d' (Section 'Topological sensing'). The DFT HOMO/LUMO localization in Fig. 3(a) is consistent with a TJS, but it does not distinguish a topologically protected interface mode from an ordinary confined state of the A60 fragment at the same energy. Moreover, no trivial control junction with the same local bonding but Z2=0 is presented, so the enhanced adsorption (-0.66 eV versus -0.27/-0.30 eV) and the twofold current increase could be caused by the abrupt interface chemistry, under-coordinated sites, methyl termination, or local strain rather than by the bulk topological invariant. This leaves the 'route to topological chemistry' attribution underdetermined.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports a tight-binding classification of the Z2 invariant for a class of chevron-type graphene nanoribbons (A60), constructs finite junctions between topological and trivial unit cells, and verifies localized topological junction states in finite clusters. It then models a methyl-terminated mirror-symmetric double junction, computes DFT adsorption energies of NO2 at three sites, and simulates coherent transport through the device with NEGF. The central claim is that the topological junction state enhances NO2 adsorption (E_a = -0.66 eV vs -0.27 and -0.30 eV at non-junction sites) and causes a two-fold increase in the transport current, thereby proposing topological junction states as a basis for gas sensing and junction-assisted chemistry.","tokens_in":13603,"tokens_out":4022,"duration_ms":41171,"significance":"If the central attribution is correct, the paper opens a genuinely new application direction for topological junction states in graphene nanoribbons, moving beyond quantum computing and spintronics into chemical sensing and reactivity. The systematic Z2 classification of the A60 family in Table 2 is a useful contribution in itself, and the paper is commendably explicit about its computational procedures and provides raw data and code for reproduction. The DFT and transport results are plausible as reported, but the connection between the topological invariant and the adsorption/transport enhancement is not yet established with the necessary controls, so the significance of the 'topological chemistry' claim is currently prospective rather than demonstrated.","major_comments":[{"comment":"The Z2 invariant is computed only for the pristine periodic A60(2,2,4,4) and AGNR(7) unit cells, while all DFT and transport calculations are performed on the methyl-terminated finite double junction of Fig. 2. The statement 'we expect one TJS per junction similar to Figure 1d' (Topological sensing) is an assumption, not a verification: methyl termination removes atoms from the pi-network, finite size can shift states, and the close proximity of the two junctions can split or annihilate the topological modes. Please demonstrate that the in-gap states in the terminated device are indeed TJSs, for example by recomputing the invariant for the terminated unit cell or by continuously tracking the states under a termination parameter, and by checking that their number and location follow bulk-boundary correspondence. Without this step, the HOMO/LUMO localization in Fig. 3(a) could equally be interpreted as ordinary confined states of the A60 fragment.","section":"Topological sensing (Fig. 2, Fig. 3(a))"},{"comment":"The paper compares adsorption at the junction with adsorption on the plain AGNR(7) and A60(2,2,4,4) segments, but these are chemically and structurally different environments. To support the 'topological chemistry' attribution, please add a geometrically matched control junction with Z2=0 on both sides and the same local bonding and termination, and compare adsorption energy and charge transfer at the corresponding site. If the trivial control gives a comparable -0.6 eV adsorption, the conclusion that the topological character boosts reactivity would not be supported. This control is load-bearing for the paper's central claim.","section":"Topological sensing (adsorption energies, Fig. 2)"},{"comment":"The transport modeling is justified as non-self-consistent by assuming 'negligible relaxation and charge transfer' (citing Ref. 68), but the adsorption event itself transfers -0.55 e Mulliken charge to NO2. Please show that a self-consistent NEGF calculation (or at least a charging-potential correction) preserves the two-fold current increase, and provide transmission spectra at representative biases to identify the resonant mechanism behind the increase. Without this, the sensor readout claim rests on an approximation whose validity is questionable exactly in the regime of interest.","section":"Sensor read-out (Fig. 4, text near 'non-self-consistent transport modeling')"}],"minor_comments":[{"comment":"There are typos: 'adsoption' should be 'adsorption' in the Fig. 3 caption, and 'N-dopped' should be 'N-doped' in the text below Fig. 4.","section":"Fig. 3 caption and text below Fig. 4"},{"comment":"The header note says 'Ligthgray' instead of 'Light gray'; the footnotes a-d would be clearer as conventional table footnotes rather than as inline parenthetical markers.","section":"Table 2"},{"comment":"The two deposit identifiers for the supporting material (zenodo.15209274 in Ref. 52 and zenodo.15672059 in the Supporting Information section) should be reconciled so that readers access a single, current dataset.","section":"References and Supplementary Information"},{"comment":"The notation '−0.43eHirshfeld charge' lacks a separator; use '−0.43 e (Hirshfeld)' for consistency with the Mulliken values.","section":"Charge-transfer reporting (Topological sensing)"},{"comment":"In the sentence reporting Eg,↓=0.51 eV, Eg,↑=1.04 eV, and Eg=1.08 eV, clarify whether Eg,↑ is computed with spin polarization and whether the spin-unpolarized gap is reported for the same relaxed geometry; the slight difference between Eg,↑ and Eg is otherwise difficult to interpret.","section":"Spin-resolved gaps (Topological sensing)"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is promising and the computational work is largely reproducible, but the central attribution of the enhanced adsorption and transport response to the topological junction state is underdetermined by the current evidence. The two needed pieces—a trivial control junction and a verification that the terminated double junction actually hosts TJSs—are feasible within the scope of the paper's methods, so I recommend major revision rather than rejection. The editors may also wish to check that the A60 classification is sufficiently self-contained given the reliance on the authors' prior work (Refs. 34, 38, 39, 41)."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: the Z2 classification of the A60 GNR family is a legitimate new result, and the idea of using topological junction states rather than edge states for sensing is genuinely novel. The paper is upfront about methods and provides code and raw data, which I appreciate. My main reservation is that the headline chemical claim—NO2 binds twice as strongly at the junction because of the TJS—is not yet separated from ordinary junction chemistry. There is no trivial (Z2=0) control junction with similar geometry, so the enhanced adsorption and the twofold current increase could just reflect under-coordinated sites, methyl termination, or local strain. The HOMO/LUMO localization in the terminated device is consistent with a TJS, but it does not distinguish a topologically protected state from an ordinary confined state at the same energy. The Z2 invariant is never recomputed for the methyl-capped unit cells; the text just says 'we expect one TJS per junction.' That is a load-bearing assumption for the sensing story.\n\nOn the plus side, the A60 classification in Table 2 is systematic and reproducible, the TB and DFT pipelines are standard, and the transport calculation is clearly described. The discussion of the 'metallic' case in Table 1 and the end-state contamination in Figure 1 are handled honestly. The recovery-time estimate is a bit hand-wavy—the attempt frequency is borrowed from CNT literature—but that is a minor issue.\n\nOverall, this is a solid computational study with a new classification and a plausible application. The central attribution of the sensing enhancement to topology is underdetermined, but that is fixable with a control calculation. I would cite it for the classification, not yet for the sensing claim, and I would send it to review with the recommendation that the referee push for a trivial control and a direct check of the TJS in the device geometry.","headline":"The A60 Z2 classification is a useful new result, but the sensing claim needs a trivial control junction before it can carry the topological weight the paper puts on it.","tokens_in":14152,"tokens_out":1866,"would_cite":true,"duration_ms":18853,"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":"Topological junction states in graphene nanoribbons act as chemically reactive sites: NO2 binds more than twice as strongly there and doubles the transport current.","keywords":["graphene nanoribbons","topological junction states","Z2 invariant","intercellular Zak phase","NO2 gas sensing","density functional theory","tight-binding model","quantum transport"],"falsifier":"Recompute the $Z_2$ invariant for the periodic ribbon whose unit cell carries the methyl termination, or solve the tight-binding model of the terminated double junction directly: if no state remains pinned near zero energy at each junction, the enhanced NO$_2$ binding and doubled current cannot be attributed to a topologically protected state. A complementary experiment is scanning tunneling spectroscopy of a methyl-terminated A60-AGNR junction, where a zero-bias conductance peak at the junction would support the claim and its absence would refute it.","tokens_in":13263,"feed_emoji":"🧪","tokens_out":10856,"duration_ms":96184,"temperature":0.7,"pith_summary":"Topological junction states are localized electronic modes that appear where a topologically nontrivial graphene nanoribbon meets a trivial one. The paper proposes using these states for chemistry and sensing instead of only for quantum computing and spintronics. It classifies a family of zigzag-shaped A60 nanoribbons by the $Z_2$ invariant, builds a mirror-symmetric double junction with a topologically nontrivial A60(2,2,4,4) segment sandwiched between trivial AGNR(7) segments, and shows by density-functional calculations that NO$_2$ binds at the junction with adsorption energy $-0.66$ eV, more than double the $-0.27$ eV and $-0.30$ eV found on the surrounding ribbon sites. Transport simulations then show that this adsorption roughly doubles the current through the device. If correct, the work establishes topological junction states as chemically reactive sites and as the active element of a gas sensor read out by electrical current.","feed_headline":"Where two nanoribbon phases meet, NO2 binds twice as strongly","feed_subtitle":"Adsorption energy nearly doubles and the sensor current doubles when NO2 lands at the junction.","key_machinery":"The $Z_2$ topological invariant computed from the intercellular Zak phase of a tight-binding $\\pi$-electron Hamiltonian. For the A60 family of zigzag-shaped nanoribbons, this invariant takes values 0 or 1 depending on the unit-cell parameters $\\ell$ and $w$; when a ribbon with $Z_2 = 1$ is joined to one with $Z_2 = 0$, bulk-boundary correspondence guarantees a localized topological junction state at the interface. The device geometry is a mirror-symmetric double junction whose dangling-bond carbon atoms are converted to methyl groups, leaving one junction state per junction, and transport through it is modeled in the low-bias regime with non-equilibrium Green's functions.","core_discovery":"The central claim is that localized topological junction states are chemically distinct: a NO$_2$ molecule binds preferentially and more strongly at the junction between a topologically nontrivial A60(2,2,4,4) nanoribbon and a trivial AGNR(7) nanoribbon. In the methyl-terminated double-junction device, the two junction states split symmetrically around the Fermi level and appear as the HOMO and LUMO; adsorption of NO$_2$ at the junction transfers roughly half an electron out of the ribbon, induces spin splitting that reduces the minority-spin gap, and results in a twofold increase in the modeled transport current. The paper reads these results as evidence that topological junction states can be exploited for sensing and junction-assisted chemistry, adding a chemical application to the quantum-computing and spintronics roles usually discussed for such states.","pith_inferences":["A natural extension not developed in the paper is to treat the methyl termination itself as part of the topological design: recomputing the $Z_2$ invariant for the terminated unit cell would show whether the sensing state is truly protected or merely a robust edge state of the finite cluster.","The same termination-and-juxtapose recipe could transfer to other one-dimensional topological platforms, such as germanene nanoribbons or engineered photonic and acoustic lattices, whose localized interface states might be turned into sensing or reaction sites.","Systematically varying the A60 parameters $\\ell$ and $w$ across the classification table could tune the energy and spatial extent of the junction states, and therefore tune adsorption energy and selectivity; the paper does not compute this design map.","The predicted twofold current increase assumes a well-matched, low-bias lead–device interface; including self-consistent transport would test whether the junction-induced current change survives realistic lead scattering, and how the sensor's sensitivity compares with the site-selective blocking seen at non-junction sites."],"forward_implications":["NO$_2$ adsorption at the topological junction is more than twice as strong as at nearby non-junction sites ($-0.66$ eV versus $-0.27$ eV and $-0.30$ eV), making the junction the reactive hot spot of the device.","Transport through the double junction is site-selective: NO$_2$ at the junction doubles the current, while adsorption on the AGNR or A60 segments blocks it, giving an electrical readout that distinguishes where the molecule binds.","The estimated room-temperature recovery time at the junction site is about 0.15 s, more than an order of magnitude faster than a recent experimental NO$_2$ sensor based on indium oxide nanoparticles, so the device could support rapid measurement cycles.","The $Z_2$ classification table for the A60 family is predictive: other parameter combinations with $Z_2 = 1$ should also host junction states when paired with trivial ribbons, extending the same sensing design.","Because the junction states are nucleophilic and radical-like, the same design principle applies to catalysis and to junction-assisted chemistry beyond gas detection."],"supporting_citations":[{"why":"Establishes topological junction states in graphene nanoribbons from the $Z_2$ invariant, the foundation for the junctions studied here.","marker":"[10]"},{"why":"Provides the finite-size junction construction and verification approach for chevron and cove-edged ribbons, extended here to A60-AGNR junctions.","marker":"[11]"},{"why":"Earlier study of electronic and adsorption properties of extended chevron ribbons, from which the A60 ribbons are revisited as topologically nontrivial.","marker":"[34]"},{"why":"Defines the A60 class of zigzag-shaped graphene nanoribbons and their structural parameters, the basis of the classification table.","marker":"[38]"},{"why":"Supplies the intercellular Zak phase formula used to evaluate the $Z_2$ invariant.","marker":"[42]"},{"why":"Connects the intercellular Zak phase to bulk-boundary correspondence, justifying the appearance of junction states.","marker":"[43]"},{"why":"Prior junction-free graphene-nanoribbon sensor whose current-blocking detection mechanism is contrasted with the current-increasing mechanism proposed here.","marker":"[69]"},{"why":"Supplies the attempt frequency used in the recovery-time estimate for the NO2 sensor.","marker":"[72]"}],"fun_headline_variants":["Topological junction states double NO2 sensing response","Graphene nanoribbon junctions: a new sensing platform for NO2","Junction topology enhances NO2 binding and current in nanoribbons","Where ribbons meet, NO2 sticks harder: topological sensing","Topological chemistry: nanoribbon junctions boost NO2 detection"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The sensing and transport results assume that the topological junction state survives after the device's dangling carbon bonds are converted to methyl groups, even though the $Z_2$ classification is computed for pristine periodic ribbons and the terminated device is only assumed, not proved, to retain one junction state per junction.","fun_headline_variants_meta":{"raw":{"variants":["Topological junction states double NO2 sensing response","Graphene nanoribbon junctions: a new sensing platform for NO2","Junction topology enhances NO2 binding and current in nanoribbons","Where ribbons meet, NO2 sticks harder: topological sensing","Topological chemistry: nanoribbon junctions boost NO2 detection"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000321,"raw_usage":{"total_tokens":1749,"prompt_tokens":828,"completion_tokens":921,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":444,"completion_tokens_details":{"reasoning_tokens":832}},"tokens_in":444,"tokens_out":921,"duration_ms":8533,"temperature":1.0,"reasoning_tokens":832,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T05:08:18.527287+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Recompute the $Z_2$ invariant for the periodic ribbon whose unit cell carries the methyl termination, or solve the tight-binding model of the terminated double junction directly: if no state remains pinned near zero energy at each junction, the enhanced NO$_2$ binding and doubled current cannot be attributed to a topologically protected state. A complementary experiment is scanning tunneling spectroscopy of a methyl-terminated A60-AGNR junction, where a zero-bias conductance peak at the junction would support the claim and its absence would refute it.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes topological junction states in graphene nanoribbons from the $Z_2$ invariant, the foundation for the junctions studied here."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the finite-size junction construction and verification approach for chevron and cove-edged ribbons, extended here to A60-AGNR junctions."},{"cited_title":"E lectronic and adsorption properties of extended chevron and cove-edged graphene nanoribbons","cited_arxiv_id":null,"evidence_quote":"Earlier study of electronic and adsorption properties of extended chevron ribbons, from which the A60 ribbons are revisited as topologically nontrivial."},{"cited_title":"A.; Batrakov, K","cited_arxiv_id":null,"evidence_quote":"Defines the A60 class of zigzag-shaped graphene nanoribbons and their structural parameters, the basis of the classification table."},{"cited_title":"N.; Car, R.; Resta, R","cited_arxiv_id":null,"evidence_quote":"Supplies the intercellular Zak phase formula used to evaluate the $Z_2$ invariant."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Connects the intercellular Zak phase to bulk-boundary correspondence, justifying the appearance of junction states."},{"cited_title":"An ultra-sensitive gas nanosensor based on asymmetric dual-gate graphene nanoribbon field-effect transistor: proposal and investigation","cited_arxiv_id":null,"evidence_quote":"Prior junction-free graphene-nanoribbon sensor whose current-blocking detection mechanism is contrasted with the current-increasing mechanism proposed here."},{"cited_title":"Ab initio study of CNT NO _2 gas sensor","cited_arxiv_id":null,"evidence_quote":"Supplies the attempt frequency used in the recovery-time estimate for the NO2 sensor."}],"review_version":1}