{"id":"6e5c41b6-45df-4ef0-8710-2d2d784e126e","arxiv_id":"2505.08914","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Flat, proximitized gold films can be grown on V(100), Nb(110), and Nb(100), and they preserve both the superconducting gap and the magnetic state of adsorbed FeTPP-Cl molecules.","lead":"This paper reports recipes for growing atomically flat gold films on superconducting vanadium and niobium crystals, and shows that these films inherit superconductivity through the proximity effect and can host magnetic molecules that keep their spin. The work gives researchers a materials platform for studying the interplay of superconductivity, magnetism, and spin-orbit coupling at surfaces.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"dGSJ and hard-gap interpretation assumes ballistic Au film, but the same films are alloyed and the paper concedes diffusion reduces the ballistic proximity effect.","rationale":"The reader's verdict is CONDITIONAL with the ballistic assumption as the weakest point; I agree. My independent reading confirms that the paper's own observation of intermixing undercuts the clean-limit assumption for the very films used for the dGSJ claim. The strongest independent evidence—the thickness-dependent gap on corrugated Nb(100) films measured with a normal tip (Fig. 5b,c) and the YSR spectra with a Nb tip—still supports the existence of a proximitized gap and magnetic-molecule integrity. However, the 'hard gap' and dGSJ resonances are presented as a central characterization, and they rely on an untested ballistic hypothesis. I also note two secondary overclaims: (1) the abstract and conclusion generalize 'oxygen-free' to all films, but O 1s XPS is only shown for Au/V(100); (2) the conclusion says films can be grown 'without the need of extensively cleaning their surface,' yet the Methods describe repeated sputtering–annealing cycles up to 1000 °C and flashes to 1600 °C. These do not change the physics verdict but support keeping the decision CONDITIONAL pending a normal-tip check of the hard gap and a diffusive-model analysis. The concern is concrete and testable, so no change in verdict is warranted.","tokens_in":17180,"tokens_out":6259,"duration_ms":62713,"concrete_test":"Record dI/dV spectra with a normal-metal tip (e.g., W or PtIr) on the same flat, annealed films used in Fig. 4—2.5 ML Au/V(100) and 6.5 ML Au/Nb(110) R1—at 1.1 K, without changing any other preparation. If the normal-tip spectrum shows strictly zero conductance at zero bias, the hard gap is genuine and the dGSJ fit in Eq. (1) is relevant. If it shows a soft gap, finite subgap conductance, or a Dynes-like lineshape, then the 'hard gap' reported in Fig. 4 is a superconducting-tip artifact, and the dGSJ assignment must be replaced by a diffusive proximity analysis; this would weaken the paper's central claim to 'proximitized gap present' without the dGSJ/hard-gap interpretation.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper interprets the zero-bias conductance suppression and the peaks at ±0.73 meV (Au/V(100)) and ±1.22 meV (Au/Nb(110)) as a hard proximity gap with dGSJ resonances described by Eq. (1). This interpretation presupposes the ballistic regime: the Au layer thickness L must be smaller than the quasiparticle mean free path l, and Andreev bound states must form in a clean N layer. The same films, however, are annealed at 450–650 °C and are partially alloyed: XPS and STM show V/Nb atoms diffused into the Au film, and Sec. III explicitly states that bulk-atom diffusion 'unequivocally reduces the ballistic proximity effect in the films.' No measurement of l, impurity concentration, or residual resistivity is given, so the ballistic assumption is unsupported for the exact films of Fig. 4. If l < L, the spectra should be analyzed with a diffusive proximity model (Usadel), and the 'hard gap' may be an artifact of the superconducting tip's own density of states rather than a zero sample DoS. The load-bearing condition for the central claim—that these films show a hard gap with dGSJ resonances close to the bulk value—is therefore not established.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports a systematic growth study of Au thin films on V(100), Nb(110), and Nb(100) using STM and XPS, with additional STS characterization of proximity-induced superconductivity and of FeTPP-Cl molecules deposited on the Au/Nb(110) films. The authors show that flat films can be obtained after post-annealing, that annealing causes intermixing with substrate atoms, and that the films exhibit a superconducting gap attributed to the proximity effect, interpreted with de Gennes-Saint-James resonances. They further report an inelastic spin excitation for FeTPP-Cl and Yu-Shiba-Rusinov states for a dechlorinated species, and conclude that these gold films are promising platforms for molecular magnetism and on-surface synthesis without extensive substrate cleaning.","tokens_in":17422,"tokens_out":8099,"duration_ms":84938,"significance":"If the central claims hold, this is a useful experimental advance: it expands the available platforms for proximity-coupled thin films beyond the well-studied Ag/Nb(110) system, and demonstrates a route to adsorb intact magnetic molecules on a superconducting gold surface. The paper contains a substantial body of new experimental data—growth morphologies across three substrate orientations, XPS identification of intermixing, STS spectra as a function of field and thickness, and molecular spectroscopy—that will be of interest to the condensed-matter and molecular-nanoscience communities. The use of standard theoretical models (Yu-Shiba-Rusinov, de Gennes-Saint-James) is appropriate and the paper is written clearly overall. The main weakness is that a load-bearing assumption, the ballistic proximity regime, is not quantitatively supported for the very films whose spectra are interpreted with Eq. (1).","major_comments":[{"comment":"The dGSJ interpretation of Fig. 4b,d presupposes a ballistic normal layer with thickness L smaller than the quasiparticle mean free path l. The manuscript itself states that diffused V/Nb atoms 'induce new quasiparticle scattering processes that reduce the electronic mean free path' and later concludes that diffusion 'unequivocally reduces the ballistic proximity effect in the films.' Yet no estimate of l, impurity concentration, or residual resistivity is given for the films measured in Fig. 4, whose STM images show abundant atomic-scale defects (Fig. 4a,c). If the films are diffusive, Eq. (1) does not describe the spectra, and the observed zero-conductance gap would have to be analyzed with a diffusive model (e.g., the Usadel equation), which can also produce a minigap for sufficiently thin layers. The central claim—that these films exhibit a hard proximity gap with dGSJ resonances close to the bulk value—is therefore not established without either a quantitative justification of L < l or a diffusive-model analysis.","section":"Sec. III, Eq. (1) and the paragraph beginning 'Having resolved the properties...'"},{"comment":"The abstract claims that gold films 'can grow flat and oxygen-free when annealed to sufficiently high temperatures,' and the conclusion states that superconducting gold films can be grown 'without the need of extensively cleaning their surface.' However, the oxygen-free claim is demonstrated only for Au/V(100), via the O 1s XPS data in Fig. 1k inset. No XPS data are presented for Nb(110) or Nb(100); the starting Nb surfaces show NbO nanocrystals (Fig. 2b) and a Nb-(3x1)-O reconstruction (Fig. 3b), and the text leaves the alloying on Nb(110) as 'very probable' based on STM alone. The generalization to all three surfaces is therefore unsupported. Additionally, the 'without extensive cleaning' phrasing understates the elaborate preparation used (sputtering, annealing, high-temperature flashes; Sec. VI). Please restrict the oxygen-free claim to V(100) or provide supporting XPS for the Nb substrates, and qualify the statement about cleaning.","section":"Abstract and Sec. IV (Conclusion); XPS data in Sec. II"},{"comment":"The conclusion that superconductivity extends further on corrugated Nb(100) films than on flat annealed V(100) films, used to support the statement that alloying reduces the ballistic proximity effect, is based on a questionable comparison. The horizontal axis in Fig. 5c is 'relative Au film apparent height' from STM on a strongly corrugated film (line profile in the inset shows >6 nm changes), not an independently calibrated film thickness; the film was annealed at 550°C and exhibits 3D facets. Comparing these data with the flat annealed V(100) films from Ref. [10] mixes different morphologies and measurement conditions, and no error bars are provided for either dataset. The quantitative claim of 'about 10% of gap reduction per nanometer' and the conclusion about films thinner than 10 MLs would need a calibrated thickness scale and error analysis to be made rigorous.","section":"Sec. III, Fig. 5 and the thickness-dependence paragraph"}],"minor_comments":[{"comment":"The phrase 'Yu-Shiva-Rusinov states' contains a misspelling of Shiba; please change to 'Yu-Shiba-Rusinov states.'","section":"Abstract"},{"comment":"The caption lists small-scale images as corresponding to '(a,c,e,g,f)'; the last entry appears to be a typo and should likely be '(a,c,e,g,i)'.","section":"Fig. 1 caption"},{"comment":"In the text 'From the set of spectra shown in Figs. 4b and Figs. 4d,' the phrase should be simplified to 'Figs. 4b and 4d.'","section":"Sec. III"},{"comment":"There are minor typographical errors: 'acknoweledges' in Sec. V, 'inhomogenious' in Sec. II, and 'magnetic fields applied is perpendicular' in Sec. VI; please proofread for such issues.","section":"Sec. VI (Methods)"},{"comment":"The statement that dGSJ states lie at ±0.73 meV and ±1.22 meV could be misread as the raw bias positions of the peaks; please clarify that these energies are obtained after subtracting the superconducting tip gap, as implied by the surrounding text.","section":"Sec. III, discussion of Fig. 4"}],"recommendation":"major_revision","confidential_remarks":"The paper contains a substantial amount of new experimental data and is likely suitable for a journal like Physical Review Materials or similar, but the interpretation overreaches in two respects: the ballistic-limit treatment is not quantitatively supported for the alloyed films, and the 'oxygen-free' claim is based on XPS data for V(100) only. The self-citations to Refs. [10] and [67] are appropriate as baselines rather than circular inputs, and the YSR analysis is consistent with the literature. I recommend major revision with a request for a mean-free-path estimate or diffusive-model fit, and for qualification of the oxygen-free statement to the substrate where it is actually measured."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThe short version: the growth recipes for Au on V(100), Nb(110) and Nb(100) are new, believable, and will be used by anyone working on proximity-coupled molecular magnetism. The YSR spectra on FeTPP-Cl on Au/Nb(110) are the strongest part — the spin survives on R1, dechlorination produces two clean species, and the chiral one sits near the quantum phase transition. That is a genuinely nice observation.\n\nWhat is new: the Nb(110) surface yields three reconstructions (R1-R3), and only R1 is chemically benign; this is the first report of YSR states on Au/Nb(110). The XPS on Au/V(100) is careful: oxygen disappears and Au-V alloy forms after annealing, with a clear thickness dependence of the alloying temperature. The thickness-dependent gap decay on corrugated Nb(100) films is a good experiment, even if the diffusion explanation is not directly verified.\n\nWhere it gets soft. First, the abstract claims 'gold films can grow flat and oxygen-free' on all surfaces, but the oxygen-free XPS evidence is only shown for V(100). For Nb, the films are grown on oxygen-reconstructed surfaces and no XPS is reported on the Au/Nb films themselves. That overgeneralization should be fixed.\n\nThe bigger issue is the dGSJ analysis. Equation (1) is the clean-limit (ballistic) expression, but the paper's own XPS and STM show the films are intermixed, and Sec. III explicitly says that diffused V/Nb atoms 'unequivocally reduce the ballistic proximity effect.' No mean free path or residual resistivity is provided. On top of that, a superconducting tip gives a zero-bias dip even for a normal-metal sample, so the hard-gap claim does not follow from the zero conductance. The peaks at ±0.73/±1.22 meV may be dGSJ resonances, but the evidence on these specific alloyed films does not force that assignment. This is fixable by fitting with a diffusive Usadel model, or at least by softening the 'hard gap' language and presenting the spectra as proximity-induced with unclear subgap structure. The lack of error bars or multiple spectra on the gap values is a minor issue for a growth paper.\n\nWho this is for: experimental groups studying molecular spins on superconductors and possibly topological superconductivity on high spin-orbit platforms. They will get real value from the recipes. I would send this to peer review, but with a request for major revision: address the ballistic/diffusive inconsistency, temper the oxygen-free claim, and provide a direct check of the gap with a normal tip (they already have such data in Fig. 5). The YSR part can stand as is.\n\nRecommendation: engage with it, but push back on the STS interpretation.\n\nBest,\n[You]","headline":"A useful growth paper with a strong YSR result, but the dGSJ/hard-gap interpretation is overstretched given the alloyed films.","tokens_in":18037,"tokens_out":5782,"would_cite":true,"duration_ms":56565,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper establishes that flat, oxygen-free gold films grown on V(100), Nb(110), and Nb(100) develop a hard, proximity-induced superconducting gap that stays above half the bulk gap for films thinner than ten monolayers, even though the…","keywords":["proximity effect","gold thin films","vanadium","niobium","de Gennes-Saint-James resonances","Yu-Shiba-Rusinov states","scanning tunneling spectroscopy","epitaxial growth"],"falsifier":"Measure the same films with a normal-metal tip and compare the sub-gap conductance against the dirty-limit prediction: under the ballistic assumption, the zero-bias conductance should stay at zero for film thicknesses well below the mean free path, whereas a diffusive film would show a finite in-gap density of states that grows with thickness; a second check is to track the dGSJ resonance energy as a function of local terrace height on a wedged film and see whether it follows Eq. (1) with L equal to the STM-measured thickness.","tokens_in":16980,"feed_emoji":"🧲","tokens_out":5173,"duration_ms":52512,"temperature":0.7,"pith_summary":"This paper establishes that flat, oxygen-free gold films can be grown on three bulk superconducting surfaces—V(100), Nb(110), and Nb(100)—by annealing after room-temperature deposition, even though the surfaces carry oxygen reconstructions and the annealing draws substrate atoms into the film. The authors show by scanning tunneling spectroscopy that these intermixed films nevertheless develop a proximity-induced superconducting gap that is hard, meaning zero conductance inside the gap, and closer to the bulk gap than to half of it for films thinner than 10 monolayers. On the flat R1 termination of Au/Nb(110), the magnetic molecule FeTPP-Cl survives adsorption with its spin intact, evidenced by spin-excitation peaks and by Yu-Shiba-Rusinov bound states at zero energy for a dechlorinated species. If correct, the work means that proximitized gold films can serve as a practical platform for combining molecular magnetism, on-surface synthesis, and superconductivity without demanding pristine, extensively cleaned superconductor surfaces.","feed_headline":"Gold films turn superconducting on V and Nb surfaces","feed_subtitle":"Flat annealed gold keeps a hard gap above half the bulk value and hosts magnetic molecules intact.","key_machinery":"The load-bearing object is the superconducting thin film itself as a ballistic proximity junction: a normal-metal gold layer of thickness L on a bulk superconductor, with quasiparticles reflected at the vacuum interface. The argument runs through the de Gennes-Saint-James resonance condition $2L\\epsilon/(\\xi \\Delta)=n\\pi+\\arccos(\\epsilon/\\Delta)$, which predicts one sub-gap bound state for $L\\ll\\xi$; the pair of conductance peaks flanking the gap is read as that resonance. STS with superconducting V or Nb tips supplies the energy resolution to see the hard gap and the resonance positions, while XPS establishes the intermixing that the argument must absorb. Yu-Shiba-Rusinov states, computed for a magnetic impurity on a proximitized gold film, serve as the magnetic probe that certifies the molecules' spin survives.","core_discovery":"The paper claims that superconducting proximity survives, and even dominates, in gold films that are deliberately grown on contaminated, reconstructed V and Nb surfaces. After annealing, the films are flat and oxygen-free but contain diffused V or Nb atoms; XPS shows an Au-V alloy on V(100) at high temperature. Despite this, STS on Au/V(100) and Au/Nb(110) shows a zero-conductance gap whose coherence-peak resonances sit at about 96% and 80% of the respective bulk gaps, which the authors attribute to de Gennes-Saint-James bound states confined between the superconductor and the vacuum. On Au/Nb(100), the gap decays by roughly 10% per nanometer of added film thickness, and the authors conclude that diffused substrate atoms reduce the ballistic proximity length. Finally, the paper asserts that FeTPP-Cl on the R1 phase of Au/Nb(110) retains its S=5/2 moment and that a chiral dechlorinated species sits at a quantum phase transition from strong to weak coupling, evidenced by YSR states pinned at the tip gap.","pith_inferences":["Editor's extension: if the hard gap survives in these partly alloyed films, the diffused V or Nb impurities may scatter quasiparticles less than their chemical presence suggests, and a direct comparison of gap hardness in deliberately alloyed versus pure films would settle how much disorder the ballistic proximity effect tolerates.","Editor's extension: the YSR states pinned at the tip-gap energy suggest the chiral FeTPP species can be tuned through the singlet-doublet quantum phase transition by tip force; a lower-temperature spectroscopic map could locate the degeneracy point and test whether the coupling can be pushed across it.","Editor's extension: the recipe of accepting intermixing could extend to other noble metals deposited on reconstructed niobium or vanadium surfaces, with the pseudomorphic silver buffer layer on Nb(110) as a natural comparison; the gap-versus-thickness slope would then become a direct probe of how much alloying degrades the ballistic proximity effect.","Editor's extension: the de Gennes-Saint-James resonance position on a corrugated film could serve as a local thickness sensor, since the resonance energy depends on the normal-metal length through Eq. (1); cross-checking STS peak positions against STM height profiles would test this directly."],"forward_implications":["Gold films on V(100) and Nb(110) thinner than 10 monolayers keep a superconducting gap greater than half the bulk value, so they can be used as superconducting platforms for molecular and device studies.","On Au/Nb(110), only the flat R1 termination preserves intact FeTPP-Cl molecules with a spin excitation, while the R2 and R3 reconstructions decompose or distort them, so surface phase selection controls molecular viability.","On Au/Nb(100), the superconducting gap drops by about 10% per nanometer of added film thickness, and less-intermixed corrugated films show slower decay, indicating that substrate-atom diffusion sets the length scale for pair correlations.","The hard gap with de Gennes-Saint-James resonances appears at critical fields of roughly 0.4 T for Au/V(100) and 0.5 T for Au/Nb(110), giving a direct measure of the induced pairing strength.","Proximitized gold films can be made without extensive cleaning of V and Nb surfaces, since flat films grow on the oxygen-reconstructed starting surfaces.","Editor's extension: if the hard gap survives in these partly alloyed films, the diffused V or Nb impurities may scatter quasiparticles less than their chemical presence suggests, and a direct comparison of gap hardness in deliberately alloyed versus pure films would settle how much disorder the ballistic proximity effect tolerates.","Editor's extension: the YSR states pinned at the tip-gap energy suggest the chiral FeTPP species can be tuned through the singlet-doublet quantum phase transition by tip force; a lower-temperature spectroscopic map could locate the degeneracy point and test whether the coupling can be pushed across it.","Editor's extension: the recipe of accepting intermixing could extend to other noble metals deposited on reconstructed niobium or vanadium surfaces, with the pseudomorphic silver buffer layer on Nb(110) as a natural comparison; the gap-versus-thickness slope would then become a direct probe of how much alloying degrades the ballistic proximity effect."],"supporting_citations":[{"why":"Reports proximitized Au films on V(100) and molecular spin behavior on them, providing the baseline this work extends to niobium surfaces.","marker":"[10]"},{"why":"Shows pure silver islands on Nb(110) with a proximitized gap close to bulk niobium, used as a comparison for the faster gap decay attributed to alloying.","marker":"[40]"},{"why":"Provides the de Gennes-Saint-James resonance condition used to assign the sub-gap conductance peaks in the proximitized films.","marker":"[65]"},{"why":"Supplies the theoretical framework for Yu-Shiba-Rusinov states on proximitized gold films, used to interpret the molecular bound states.","marker":"[66]"},{"why":"Demonstrates YSR states and tip-force tuning on Au/V(100), the benchmark for the chiral FeTPP spectra on Au/Nb(110).","marker":"[67]"},{"why":"Reports Au-V alloy formation detected by XPS, used to interpret the chemical shifts observed after annealing Au on V(100).","marker":"[26]"},{"why":"Establishes that a clean Nb(110) surface requires flashing near the melting point, justifying the choice to grow gold on oxygen-reconstructed surfaces.","marker":"[32]"},{"why":"Studies Au films on V(110) in a topological proximity context, motivating the use of gold on vanadium superconductors.","marker":"[17]"}],"fun_headline_variants":["Gold films on V and Nb superconduct by proximity","Near-bulk superconducting gap in gold on V and Nb","Flat gold on V and Nb with near-bulk superconducting gaps","Superconducting gold films on V and Nb host magnetic molecules","Proximitized gold on V and Nb: superconductivity without oxide"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the annealed, partly alloyed gold films remain in the ballistic proximity regime, with film thickness shorter than the quasiparticle mean free path, so the de Gennes-Saint-James interpretation and the hard-gap conclusion apply; if the diffused vanadium or niobium atoms instead make the films diffusive, the same spectra would need a dirty-limit model and the hard-gap claim would not follow.","fun_headline_variants_meta":{"raw":{"variants":["Gold films on V and Nb superconduct by proximity","Near-bulk superconducting gap in gold on V and Nb","Flat gold on V and Nb with near-bulk superconducting gaps","Superconducting gold films on V and Nb host magnetic molecules","Proximitized gold on V and Nb: superconductivity without oxide"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00202,"raw_usage":{"total_tokens":7920,"prompt_tokens":1033,"completion_tokens":6887,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":649,"completion_tokens_details":{"reasoning_tokens":6803}},"tokens_in":649,"tokens_out":6887,"duration_ms":48314,"temperature":1.0,"reasoning_tokens":6803,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T21:45:33.445429+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the same films with a normal-metal tip and compare the sub-gap conductance against the dirty-limit prediction: under the ballistic assumption, the zero-bias conductance should stay at zero for film thicknesses well below the mean free path, whereas a diffusive film would show a finite in-gap density of states that grows with thickness; a second check is to track the dGSJ resonance energy as a function of local terrace height on a wedged film and see whether it follows Eq. (1) with L equal to the STM-measured thickness.","supporting_citations":[{"cited_title":"Vaxevani, J","cited_arxiv_id":null,"evidence_quote":"Reports proximitized Au films on V(100) and molecular spin behavior on them, providing the baseline this work extends to niobium surfaces."},{"cited_title":"Schneider, K","cited_arxiv_id":null,"evidence_quote":"Shows pure silver islands on Nb(110) with a proximitized gap close to bulk niobium, used as a comparison for the faster gap decay attributed to alloying."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the de Gennes-Saint-James resonance condition used to assign the sub-gap conductance peaks in the proximitized films."},{"cited_title":"Ortuzar, J","cited_arxiv_id":null,"evidence_quote":"Supplies the theoretical framework for Yu-Shiba-Rusinov states on proximitized gold films, used to interpret the molecular bound states."},{"cited_title":"Trivini, J","cited_arxiv_id":null,"evidence_quote":"Demonstrates YSR states and tip-force tuning on Au/V(100), the benchmark for the chiral FeTPP spectra on Au/Nb(110)."},{"cited_title":"Valla, P","cited_arxiv_id":null,"evidence_quote":"Reports Au-V alloy formation detected by XPS, used to interpret the chemical shifts observed after annealing Au on V(100)."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes that a clean Nb(110) surface requires flashing near the melting point, justifying the choice to grow gold on oxygen-reconstructed surfaces."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Studies Au films on V(110) in a topological proximity context, motivating the use of gold on vanadium superconductors."}],"review_version":1}