{"id":"f0d24bec-6930-47db-bbda-9abd26b98afd","arxiv_id":"2502.03448","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"U3Bi4Ni3 is reported to be a Kondo insulator with a metallic surface state that dominates transport below 150 K and carries roughly ten times more current than the bulk at low temperature.","lead":"A uranium-based compound, U3Bi4Ni3, conducts electricity on its surface even though its interior is an insulator, and this surface behavior works at much higher temperatures than in other Kondo insulators. The discovery gives researchers a new material platform for studying how strong electron interactions and surface conduction coexist.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The FEA decomposition is exactly determined (two resistances, two fit parameters) and assumes an infinitely thin surface layer; the extracted σs(T) and its 250 K onset are therefore model-imposed rather than independently verified.","rationale":"The paper's central claim is that U3Bi4Ni3 hosts a surface state that emerges below 250 K and dominates transport below 150 K. The most load-bearing assumption is that the transport can be decomposed into an infinitely thin 2D surface conductor in parallel with a 3D bulk, and that the FEA extraction of ρb and σs is meaningful. The Methods describe fitting two measured resistances (R12;12 and R12;34) at each temperature with two parameters (ρb, σs), using 'Electric Currents in Layered Shells' for the surface. This is an exactly determined system, so the extracted σs(T) cannot be wrong within the model—but it also cannot validate the model. Any temperature dependence in σs, including the claimed onset near 250 K, is forced by the data only under the assumed geometry. If a finite-thickness surface layer, an impurity band, or a damaged subsurface layer carries the extra conduction, the same raw data would be reproduced by different parameters and the onset and magnitude claims would change. The paper provides some independent support: the thickness-dependent resistivity in Fig. 1d shows that the low-temperature resistance decreases with thickness, and the R12;12/R12;34 Corbino data show an anomalous 'inverse gap' behavior. These observations make the existence of a high-conductivity near-surface channel plausible. The concern is not that the data are fabricated; it is that the quantitative decomposition—and especially the conclusion that the channel is 'inherently protected'—is not overdetermined by the measurements. The robustness argument rests on only two Corbino samples and one polishing series, and the paper itself notes that DFT+U does not capture the Kondo physics (SI Section III), so no microscopic calculation supports topological protection. The ARPES data in Fig. 2 show insulating EDCs without any surface-state dispersion. Thus, the conditional verdict is appropriate: the work is a valuable candidate report, but the 'surface state at high temperature' and 'inherently protected' claims require direct surface-sensitive measurements or a thicker, overdetermined transport dataset. The proposed test—successive polishing of a single Corbino sample and re-extracting σs at each thickness—would settle whether the 'surface' is genuinely 2D or a finite-depth layer. An alternative, cheaper test is to re-fit the existing data with a finite-thickness subsurface-layer model; if the fit quality is comparable, the infinite-thin-surface assumption is not uniquely determined.","tokens_in":15354,"tokens_out":9467,"duration_ms":82304,"concrete_test":"Perform the inverted-resistance measurement on a single Corbino sample while successively polishing it to several thicknesses (e.g., 210, 150, 100, and 50 µm). At each thickness, extract ρb(T) and σs(T) using the published FEA procedure. If the 2D surface-state interpretation is correct, σs(T) must be independent of thickness (up to sample-to-sample variation), while ρb(T) is unchanged; if σs(T) systematically changes with thickness, the high-conductivity channel is a finite-depth damage or subsurface layer rather than a protected surface state.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that a surface state emerges below 250 K and dominates transport below 150 K rests on the bulk/surface decomposition extracted by FEA (Methods, 'Finite Element Analysis'). The FEA uses two measured four-terminal resistances, R12;12 and R12;34, at each temperature and fits them with exactly two parameters, ρb and σs, using an infinitely thin 'Electric Currents in Layered Shells' surface channel. The system is therefore exactly determined: the temperature dependence shown for σs in Fig. 3e,h is the value required to reproduce the data under the assumed 2D-in-parallel-with-3D geometry, not an independent measurement. If the high-conductivity path is instead a finite-thickness subsurface layer (damaged layer, surface band-bending accumulation, or impurity band), the same raw resistances would be fit by a different σs(T) and a different onset; the abstract's 'below 150 K' is also sample-dependent, since S2 crosses over only near 50 K (Fig. 3i). Additionally, the inference that the state is 'inherently protected' (Section III) is based on robustness across two Corbino samples and one polishing series, and the paper's own DFT+U fails to capture the Kondo gap (SI Section III), so no microscopic calculation supports topological protection. The ARPES in Fig. 2 shows only bulk-insulating EDCs; no surface-state dispersion is presented.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript reports transport, magnetization, Hall effect, ARPES, and DFT+U results on the Kondo insulator U3Bi4Ni3. The authors infer a near-surface high-conductivity channel from thickness-dependent longitudinal resistivity and from Corbino-disk inverted-resistance measurements, and they use finite element analysis (FEA) to decompose the measured resistances into a bulk resistivity ρb(T) and a surface conductivity σs(T). They claim that a surface state emerges below 250 K and dominates transport below 150 K, with surface conductivity about one order of magnitude higher than bulk at low temperature, and they argue that the robustness of this channel indicates inherent (topological) protection. Bulk characterization, including Curie-Weiss susceptibility, activation gaps of 95 meV and 1.5 meV, anomalous Hall behavior, and resonant ARPES, is presented to establish the Kondo insulating ground state.","tokens_in":15636,"tokens_out":5340,"duration_ms":51147,"significance":"If the central claim were established, U3Bi4Ni3 would be qualitatively distinct from SmB6 and YbB12, with surface-dominated transport at temperatures far above the few-Kelvin scale of those materials; the large sheet conductivity and unusually high operating temperature would make the system attractive for devices and for studying surface-bulk competition in correlated insulators. The thickness-dependent resistivity and the inverted Corbino resistance behavior provide credible qualitative evidence for a near-surface conducting channel, and the paper usefully documents sample-to-sample variability. However, the quantitative surface/bulk decomposition is exactly determined rather than overdetermined, and the microscopic origin and protection claims are not supported by the data presented; these are load-bearing issues that need revision.","major_comments":[{"comment":"The FEA decomposition uses two measured four-terminal resistances, R12;12 and R12;34, and exactly two fitting parameters, ρb and σs, at each temperature, under the assumption of an infinitely thin 2D surface shell in parallel with the 3D bulk. The decomposition is therefore exactly determined, and the extracted σs(T) and its onset near 250 K are properties imposed by the assumed geometry rather than independently measured. If the high-conductivity path were instead a finite-thickness subsurface layer, a band-bending accumulation layer, or a damage layer, the same raw resistances would yield different σs(T) and a different onset. I request a sensitivity analysis that includes a finite-thickness surface/sub-surface layer, or an additional measurement geometry that overdetermines σs(T), before the quantitative surface-state claim is accepted.","section":"Methods, Finite Element Analysis; Fig. 3"},{"comment":"The abstract's statement that the surface state 'dominates transport properties below 150 K' is not representative of sample S2, for which the FEA ratio σb*t/σs reaches unity only near 50 K (Fig. 3i). This is not a cosmetic issue: the crossover temperature is a headline quantitative result and varies by more than a factor of three between the two samples. The claims should be stated per sample, or a well-defined aggregation with error estimates should be provided.","section":"Abstract; Fig. 3i"},{"comment":"The paper overinterprets robustness as inherent topological protection. Surface-dominated transport across polished, air-exposed surfaces and two Corbino samples is also consistent with non-topological mechanisms such as surface accumulation, polishing damage, or an impurity band. The manuscript itself states that DFT+U does not capture the Kondo gap (SI Section III), and the ARPES data in Fig. 2 show only bulk EDCs with no surface-state dispersion. The abstract's 'inherently protected' therefore overstates the evidence; this wording should be removed or explicitly qualified unless direct evidence, such as surface-state ARPES dispersion, quasiparticle interference, or a symmetry/topology calculation, is added.","section":"Section III, Discussions; SI Section III; Fig. 2"},{"comment":"The claim that surface conductivity is 'about one order of magnitude higher' than bulk is sample-dependent: at base temperature the FEA gives σs roughly 5 times the bulk value for S1 (Fig. 3f) and roughly 10 times for S2 (Fig. 3i). The text acknowledges this variation, but the abstract and later discussion present a single order-of-magnitude statement. The quantitative claim should be qualified per sample or accompanied by an uncertainty estimate that reflects the spread.","section":"Section II; Fig. 3f and 3i"}],"minor_comments":[{"comment":"The main text and Fig. 2 caption state the ARPES measurement temperature as T = 20 K, while the Methods section states the sample temperature was 6 K; please reconcile these values.","section":"Fig. 2 caption vs. Methods, ARPES"},{"comment":"The inset equation appears as 'ρ(400 K)/ρ = (2σs/σ400 K) t−1 + σb/σ400 K' and is missing a subscript on the denominator of the left-hand side; please define all symbols and verify the dimensional consistency of the displayed formula.","section":"Fig. 1d inset"},{"comment":"The text says the R12;12 configuration 'has similar behavior to ρxx(T)', but R12;12 is a resistance while ρxx is a resistivity; please specify whether the comparison is made to the resistance of a standard bar or to the resistivity, and explain how the different geometries are normalized.","section":"Section II, Corbino-disk measurements"},{"comment":"The phrase 'insulator-metal crossover' in connection with R12;34 is ambiguous because the bulk remains insulating throughout the measured range; please specify that 'metal' refers to the surface-dominated transport channel.","section":"Section II"},{"comment":"The FEA fitting procedure is described as first fitting the ratio R12;12/R12;34 and then using R12;12 (or R12;34) individually; please clarify whether both experimental resistances are fit simultaneously, and describe how uncertainties in the measured resistances propagate into the extracted ρb(T) and σs(T).","section":"Methods, Finite Element Analysis"},{"comment":"SI Figure 2 shows a carrier-type crossover near 150-200 K that is said to align with the surface-state onset; since the Hall measurement is primarily a bulk probe, please clarify whether this crossover is expected to be affected by surface conduction and why the two-band bulk fit remains valid in the surface-dominated regime.","section":"SI, Magneto-transport fittings"}],"recommendation":"major_revision","confidential_remarks":"The qualitative finding of a near-surface conducting channel is well supported by the thickness dependence, and the paper would be an important contribution if the quantitative decomposition can be made robust. The main liability is the confidence placed in the exactly determined FEA model and in the topological-protection interpretation; both are fixable through additional modeling or careful reframing. I would not reject on the current evidence, but the abstract and Section III need to be brought in line with what the data actually constrain."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: the paper reports transport evidence for a surface-near conducting channel in the Kondo insulator U3Bi4Ni3 that sets in around 200–250 K, well above the few-kelvin regime of SmB6 and YbB12. If the decomposition holds, that is a genuinely new energy scale. My read: the qualitative case for surface-near transport is solid, but the quantitative extraction of σs(T) and the 'inherently protected' conclusion go beyond what the data support.\n\nWhat the paper does well: the thickness-dependent resistivity is a clean probe; below ~150 K, resistivity drops as the sample is thinned, and the slope/intercept analysis is consistent with a conducting sheath. The Corbino inverted-resistance measurements show the expected behavior for surface-dominated transport. The FEA is done with realistic geometry, and the authors are honest about its limitations: the SI admits that DFT+U cannot capture the Kondo gap and that a FIB-milled device shows a hump rather than a plateau. That is good reporting.\n\nThe soft spot is exactly what the stress-test flags. The FEA fits two measured resistances with two parameters (ρb and σs) under an infinitely thin surface-layer assumption. That system is exactly determined, so the temperature dependence of σs—including the 250 K onset—is the value required to reproduce the data in that geometry, not an independent measurement. A finite-thickness subsurface layer, an impurity band, or a polishing damage layer would change the extracted σs and the onset. The abstract's 'below 150 K' is also sample-dependent: sample 2's crossover is near 50 K. The 'inherently protected' claim rests on robustness across two Corbino samples and one polishing series; there is no surface-sensitive spectroscopy or topology calculation. The ARPES shows insulating bulk EDCs but no surface-state dispersion. The authors themselves say DMFT is needed.\n\nWho is this for: people working on Kondo insulators and surface transport. It identifies a new candidate with a much higher surface-transport temperature scale. It deserves peer review, but a referee should ask for raw data, uncertainty estimates on the FEA outputs, and direct surface probes before accepting the quantitative decomposition. I would not block the qualitative claim, but the protection language and specific σs values should be presented as model-dependent, not measured.","headline":"A plausible high-temperature surface transport signature in a new Kondo insulator, but the quantitative decomposition and 'inherently protected' claim are model-dependent and underdetermined.","tokens_in":16253,"tokens_out":3049,"would_cite":true,"duration_ms":26218,"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":"A Kondo insulator shows a metallic surface state that dominates transport below 150 K.","keywords":["Kondo insulator","surface state","U3Bi4Ni3","uranium intermetallic","Corbino-disk transport","finite element analysis","strongly correlated electrons","topological insulator"],"falsifier":"A temperature-dependent ARPES or STM study between 100 and 250 K: if no metallic surface band appears above the Kondo temperature while the transport plateau persists, the surface-state onset claim fails; alternatively, a thickness series extending to micron-scale samples that deviates from the 1/t scaling predicted by the two-channel model would indicate a finite-depth conducting layer.","tokens_in":1813,"feed_emoji":"","tokens_out":4564,"duration_ms":89497,"temperature":0.7,"pith_summary":"This paper reports that the Kondo insulator U3Bi4Ni3 conducts electricity through a metallic surface state at temperatures far above any previously studied Kondo insulator. The authors show that a surface conduction channel emerges below 250 K and takes over transport below 150 K, whereas in SmB6 and YbB12 the surface dominates only below a few kelvin. At low temperature the surface conductivity is roughly an order of magnitude higher than the bulk. If the claim holds, U3Bi4Ni3 becomes the first Kondo insulator in which robust surface conduction is accessible at practical temperatures, opening a new material platform for studying topological surface states in strongly correlated systems and for devices based on them.","feed_headline":"Surface state dominates transport in a Kondo insulator at 150 K","feed_subtitle":"Surface conductivity runs ten times the bulk, and it appears above the Kondo temperature, unlike SmB6 and YbB12.","key_machinery":"The argument rests on the inverted-resistance measurement in a Corbino-disk geometry, in which electrode 2 acts as a two-dimensional Faraday cage and two four-terminal resistances, R12;12 and R12;34, are measured. In the surface-dominated regime R12;12 is proportional to the surface resistivity, while R12;34 follows R12;34 = C t (rho_s)^2 / rho_b, an inverse-gap scaling that makes bulk and surface contributions distinguishable. Finite-element simulations of the exact electrode geometry use the bulk resistivity rho_b and surface conductivity sigma_s as fitting parameters to reproduce both measured resistances at each temperature, yielding separate temperature dependences for the two channels. Thickness-dependent resistivity provides an independent qualitative check: above 200 K the resistivity is thickness independent, indicating bulk transport, whereas at low temperature it falls as the sample is thinned, indicating surface transport.","core_discovery":"The central discovery is that U3Bi4Ni3, a uranium-based Kondo insulator with reported spin and charge gaps near 19 meV and 72-95 meV, possesses a robust two-dimensional metallic surface state. Transport on successively polished crystals shows the resistivity becoming thickness-dependent below about 200 K, and Corbino-disk inverted-resistance measurements combined with finite-element analysis separate bulk and surface channels: the surface conductivity rises from about 250 K and becomes nearly temperature independent below 150 K, while the bulk remains activated. At base temperature the surface conductivity exceeds the bulk by roughly an order of magnitude, depending on sample. Because the surface state appears above the Kondo temperature TK approximately 100 K and survives air exposure, polishing, and sample-to-sample variation, the authors conclude it is inherently protected; they argue the most plausible origin is topological band inversion, with Kondo hybridization renormalizing the gap downward without closing it, so the topology is preserved throughout.","pith_inferences":["The onset of the surface state above TK suggests the metallic surface band may exist before Kondo coherence develops; temperature-dependent ARPES or STM between 100 and 250 K, which the paper does not report, would directly test this.","If the high-conductivity channel is a finite-thickness subsurface layer rather than an infinitely thin two-dimensional sheet, the extracted sigma_s and the 'inherently protected' interpretation would need revision; polishing samples to micrometer scale or using local probes would distinguish the pictures.","The very high sheet conductivity in a strongly correlated insulator raises the possibility of unconventional quantum oscillations in U3Bi4Ni3 at temperatures far more accessible than in SmB6.","The paper's DFT+U results fail to capture the Kondo physics, so a dynamical mean-field calculation is the natural next step; a computed topological invariant would turn the 'inherently protected' claim into a concrete, checkable prediction."],"forward_implications":["U3Bi4Ni3 becomes a third Kondo insulator platform, alongside SmB6 and YbB12, in which surface-state and quantum-oscillation questions can be studied at much higher temperatures.","Surface-dominated conduction below 150 K means bulk and surface transport can be separated over a wide temperature window without dilution-refrigerator temperatures.","The measured sheet conductivity at 5 K, 0.1-0.5 1/ohm, is an order of magnitude larger than in SmB6 and FeSi and comparable to Bi2Se3, suggesting high mobility in a strongly correlated two-dimensional system.","Because the gap renormalizes without closing through the Kondo temperature, a topological surface state would retain its topological character across the Kondo crossover.","The robust, air-stable surface conduction is a candidate building block for heterostructures aimed at quantum Hall states or proximity-induced Majorana physics."],"supporting_citations":[{"why":"Supplies the synthesis, crystal structure, and initial susceptibility evidence that U3Bi4Ni3 is a Kondo insulator candidate.","marker":"[41]"},{"why":"Provides NMR/NQR evidence for a spin gap and Kondo screening below 100 K, grounding the Kondo insulating ground state.","marker":"[42]"},{"why":"Introduces the inverted-resistance Corbino-disk method and the formula R12;34 = C t (rho_s)^2 / rho_b used to decouple bulk and surface conductivities.","marker":"[54]"},{"why":"Establishes the inverted-resistance technique as reliable in SmB6 and demonstrates a disorder-protected transport gap, serving as the methodological and comparative benchmark.","marker":"[15]"},{"why":"Documents the low-temperature resistivity plateau in SmB6, the standard signature of surface-dominated transport that this paper extends to higher temperatures.","marker":"[29]"},{"why":"Reports topological surface conduction in YbB12, the other comparison Kondo insulator whose few-kelvin surface regime U3Bi4Ni3 surpasses.","marker":"[30]"},{"why":"Provides FeSi and FeSb2 sheet-conductivity data and further validation of inverted-resistance transport in strongly correlated insulators.","marker":"[55]"},{"why":"Supplies the Bi2Se3 sheet-conductivity comparison used to argue that U3Bi4Ni3's surface conductance is unusually high.","marker":"[59]"}],"fun_headline_variants":["Uranium Kondo insulator reveals robust surface state above 150 K","Surface conduction dominates transport in U3Bi4Ni3 below 150 K","Robust surface state appears above 150 K in Kondo insulator","Surface state in U3Bi4Ni3 outshines bulk tenfold","U3Bi4Ni3: surface state wins over bulk tenfold"],"cache_read_input_tokens":18304,"weakest_assumption_plain":"The bulk-vs-surface decomposition assumes the surface is an infinitely thin two-dimensional conductor in parallel with a uniform three-dimensional bulk, and the finite-element fit uses that two-channel model to extract sigma_s and rho_b from two measured resistances; if the conductive layer has finite thickness, or is an impurity band or damage layer, the extracted values, the 250 K onset, and the protection conclusion would all need revision.","fun_headline_variants_meta":{"raw":{"variants":["Uranium Kondo insulator reveals robust surface state above 150 K","Surface conduction dominates transport in U3Bi4Ni3 below 150 K","Robust surface state appears above 150 K in Kondo insulator","Surface state in U3Bi4Ni3 outshines bulk tenfold","U3Bi4Ni3: surface state wins over bulk tenfold"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000841,"raw_usage":{"total_tokens":3656,"prompt_tokens":930,"completion_tokens":2726,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":546,"completion_tokens_details":{"reasoning_tokens":2629}},"tokens_in":546,"tokens_out":2726,"duration_ms":18096,"temperature":1.0,"reasoning_tokens":2629,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-09T04:42:34.260089+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A temperature-dependent ARPES or STM study between 100 and 250 K: if no metallic surface band appears above the Kondo temperature while the transport plateau persists, the surface-state onset claim fails; alternatively, a thickness series extending to micron-scale samples that deviates from the 1/t scaling predicted by the two-channel model would indicate a finite-depth conducting layer.","supporting_citations":[{"cited_title":"Klimczuk, H.-o","cited_arxiv_id":null,"evidence_quote":"Supplies the synthesis, crystal structure, and initial susceptibility evidence that U3Bi4Ni3 is a Kondo insulator candidate."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides NMR/NQR evidence for a spin gap and Kondo screening below 100 K, grounding the Kondo insulating ground state."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces the inverted-resistance Corbino-disk method and the formula R12;34 = C t (rho_s)^2 / rho_b used to decouple bulk and surface conductivities."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes the inverted-resistance technique as reliable in SmB6 and demonstrates a disorder-protected transport gap, serving as the methodological and comparative benchmark."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Documents the low-temperature resistivity plateau in SmB6, the standard signature of surface-dominated transport that this paper extends to higher temperatures."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Reports topological surface conduction in YbB12, the other comparison Kondo insulator whose few-kelvin surface regime U3Bi4Ni3 surpasses."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides FeSi and FeSb2 sheet-conductivity data and further validation of inverted-resistance transport in strongly correlated insulators."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the Bi2Se3 sheet-conductivity comparison used to argue that U3Bi4Ni3's surface conductance is unusually high."}],"review_version":1}