{"id":"6b39ef16-9f54-45d1-ae63-d58849f581e1","arxiv_id":"1908.03312","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Polarized optical and photoemission measurements show that RuAs opens an electronic gap along the c axis near 250 K and along the b axis near 200 K, linking the two-step metal-insulator transition to anisotropic gap formation.","lead":"The authors measured how RuAs absorbs polarized light along different crystal directions as it cools. They found that the two-step metal-to-insulator transition comes from electronic gaps opening at two different temperatures in the two directions.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The anisotropic gap-opening temperatures in Fig. 7 may be fitting artifacts: no uncertainties or model checks are reported for the Drude–Lorentz extraction that is the only evidence for onsets near 240 K versus 205 K.","rationale":"The central claim is not independently supported in the MT phase, because the PE data are too surface-sensitive near E_F to resolve a 240 K versus 205 K onset, and the resistivity data in Ref. 10 are not used to benchmark the optical gap-opening temperatures. The proposed bootstrap test directly targets whether the two onset temperatures are distinguishable given the stated measurement uncertainty. If the fit is stable under noise and extrapolation choice, the paper's claim survives; if not, the anisotropy conclusion should be downgraded to an observation of different low-temperature gap sizes only. This is a concrete, fair test, not a manufactured objection.","tokens_in":9580,"tokens_out":4644,"duration_ms":54819,"concrete_test":"Bootstrap the entire analysis chain: using the stated ±1–2% reflectivity accuracy, generate 1000 synthetic R(omega) sets with correlated noise; extrapolate below 15 meV using both the Hagen–Rubens and constant forms; run the authors' Drude–Lorentz fit with the same oscillator counts at 190, 200, 210, 230, 250, and 280 K; and record the lowest Lorentz peak energy and 0–1.2 eV centroid. If the 95% bootstrap confidence intervals for the c-axis and b-axis onset temperatures overlap, or if switching the low-energy extrapolation shifts either onset by more than about 10 K, then the claimed anisotropic gap-opening temperatures are not statistically established.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim depends on the assertion in Section III.C and Fig. 7 that the lowest Lorentzian peak energy (h-bar omega_0) and the 0–1.2 eV center of gravity (langle h-bar omega rangle) stay constant above T_MI1 on the c axis and above T_MI2 on the b axis, then rise below. This requires that the least-squares decomposition into one Drude and three/two Lorentz oscillators (Eq. in Section III.C) uniquely separates Drude spectral weight from the interband feature taken as the gap. That identification is not demonstrated. No uncertainties on fitted parameters are reported; the number of oscillators is chosen 'because of spectral shapes'; and no model comparison, residual analysis, or confidence interval is given. In the high-temperature phase the Drude term contributes strong low-energy spectral weight, so a temperature-dependent Drude damping or weight can shift the apparent position and centroid of the neighboring Lorentzian even if the interband gap edge is unchanged. Consequently, the c-axis onset at 240 ± 10 K and the b-axis onset at 205 K may be fitting-parameter trade-offs rather than real gap-opening temperatures. The Kramers–Kronig extrapolation below 15 meV (constant below 190 K, Hagen–Rubens above 210 K) is a second uncontrolled input to the centroid, which integrates down to 0 eV. The paper's claim that the gap along c opens near T_MI1 and along b opens below T_MI2 therefore rests on an unvalidated proxy.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports polarized optical conductivity, angle-integrated photoemission, and LDA band-structure calculations for RuAs, which undergoes a two-step metal-to-insulator transition at approximately 250 K (T_MI1) and 200 K (T_MI2). The authors show that the high-temperature and low-temperature electronic structures can be largely explained by LDA-based interband calculations without invoking strong electron correlations. The central new claim is that the energy gap opens anisotropically: along the c axis the gap starts to open near T_MI1, while along the b axis it opens only below T_MI2. This claim is drawn from temperature-dependent Drude-Lorentz fits to the optical conductivity and from the temperature-dependent center of gravity of the spectra, and it is used to propose that the two-step transition originates from two direction-dependent charge-ordering events.","tokens_in":9870,"tokens_out":3579,"duration_ms":39835,"significance":"If the anisotropic gap-opening claim is correct, the result is significant because it provides direct electronic-structure evidence that the two-step MIT in RuAs arises from anisotropic charge ordering, and it connects the optical response to the independently characterized structural transitions. The paper also contains a genuinely useful consistency check: the polarized optical conductivity and photoemission spectra in the HT and LT phases are compared with LDA calculations and show good mutual agreement, which is a strength. The temperature-dependent PE intensity near E_F and the decreasing Drude weight are consistent with carrier depletion across the transition. However, the central claim about different gap-opening temperatures rests on a fitting procedure for which no uncertainties, model comparisons, or robustness checks are reported, and this is the main weakness of the manuscript.","major_comments":[{"comment":"The central claim that the c-axis gap opens near 240 K and the b-axis gap near 205 K is inferred from the temperature dependence of the lowest Lorentz peak energy (ħω0) and the 0–1.2 eV center of gravity (⟨ħω⟩). These quantities are extracted from a least-squares decomposition into one Drude and either three or two Lorentz oscillators, with the number of oscillators chosen “because of spectral shapes.” No uncertainties on the fitted parameters, no residual analysis, and no alternative model comparisons are provided. Because the Drude term has strong low-energy spectral weight, temperature-dependent Drude damping or spectral weight can shift the apparent position and centroid of the neighboring Lorentzian even if the underlying interband edge is unchanged. The manuscript therefore does not rule out that the reported onsets are fitting-parameter trade-offs rather than genuine gap-opening temperatures. I request a robustness analysis: for example, fits with different oscillator counts, fits in which Drude parameters are constrained by DC conductivity, and confidence intervals on ħω0 and ⟨ħω⟩.","section":"Section III.C, Fig. 7"},{"comment":"The optical conductivity is obtained by Kramers-Kronig analysis with the reflectivity extrapolated below 15 meV using a Hagen-Rubens form at T ≥ 210 K and a constant at T ≤ 190 K. The authors state that this extrapolation does not affect the spectra around 100 meV “so much,” but no quantitative estimate is given. This matters for the central claim because the center of gravity ⟨ħω⟩ integrates the conductivity down to zero energy, and the fitted Lorentz peak energies of 0.2–0.4 eV are not far above the extrapolated region. A quantitative test of how the extracted onsets and peak energies change under alternative low-energy extrapolations is needed, as is ideally measurement to lower photon energies.","section":"Section II, Kramers-Kronig extrapolation"},{"comment":"The BCS gap functions shown in Fig. 7 are computed using assumed transition temperatures of 205 K (E ‖ bH) and 240 K (E ‖ cH), so the statement that the temperature dependence is “explained by the gap functions” is partly built into the fit and does not independently determine or validate the onset temperatures. Moreover, the data points in Fig. 7 are shown without error bars, and the assertion that ħω0 and ⟨ħω⟩ are “almost constant” above the transition and then rise below is not accompanied by any statistical test. Please report uncertainties on the extracted peak energies and centroids, and, if a BCS-like fit is used, either treat the transition temperature as a free parameter with an uncertainty or clearly label the curves as a guide based on independent resistivity/structural data.","section":"Section III.C, BCS fit and error bars"}],"minor_comments":[{"comment":"There is a typographical error in “To clarify the above inferencethe σ(ω) spectra have been fitted”; “inferencethe” should be “inference, the.”","section":"Section III.C"},{"comment":"The value 2Δ/k_B T_c ∼ 19 is presented without specifying which measured gap (b-axis or c-axis) and which transition temperature were used to compute it. The definition should be stated explicitly because the ratio is used to argue against a Peierls-like transition.","section":"Section I / Section III.C"},{"comment":"The quantity ⟨ħω⟩, called the center of gravity of the σ(ω) spectra, is not defined in the text or figure caption. Please state the definition, including the integration range and whether the Drude contribution is included.","section":"Fig. 7"},{"comment":"The comparison of measured σ(ω) with calculated interband-only spectra is informative, but the caption of Fig. 4 does not state explicitly that the calculated curves contain no Drude contribution. Adding this clarification would prevent a misleading visual comparison.","section":"Fig. 4 and surrounding text"}],"recommendation":"major_revision","confidential_remarks":"The experimental work appears carefully performed and the LDA comparison in the HT and LT phases is convincing. The main issue is the quantitative support for the central anisotropic gap-opening claim; this is fixable with additional fitting robustness checks and uncertainty reporting, so I recommend major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my read on arXiv:1908.03312. The genuinely new observation is the axis-resolved gap-opening temperature: the c-axis gap starting to open near T_MI1 (~250 K) and the b-axis gap below T_MI2 (~200 K). Everything else—valence-band photoemission, LT gap sizes, consistency with LDA—confirms an earlier picture. That new observation is the load-bearing claim, and it is not as solid as it should be.\n\nThe paper does several things well. The comparison of PE and sigma(omega) with LDA in both HT and LT phases is convincing: the peaks and spectral shapes line up, and no correlation effects are needed. The LT gap sizes (0.22 eV along b, 0.4 eV along c) match LDA, and the raw temperature-dependent spectra in Fig. 5 clearly show the gap developing in the right temperature region. This is a competently executed optical study.\n\nThe soft spot is the extraction of the onset temperatures. Section III.C fits sigma(omega) with one Drude and two or three Lorentzians, with the number of oscillators chosen 'because of spectral shapes,' and no uncertainties or model comparisons are reported. A temperature-dependent Drude width or weight can shift the fitted Lorentzian peak position even if the interband edge is fixed, and the centroid integrates the low-energy extrapolation, which changes from Hagen-Rubens to constant across the transition. So the 240±10 K and 205 K onsets in Fig. 7 could be fitting trade-offs. The BCS curves are computed with those assumed Tc values, so they cannot independently confirm the onsets. The paper partially acknowledges this by saying the spectra are 'suggested' but still presents these numbers as the main result.\n\nI'd describe the claim as plausible but unproven. It needs either a model comparison (fixed number of oscillators with a robust gap edge, e.g., spectral weight transfer in a defined window) or at least bootstrap confidence intervals on the peak positions. The tiny sample size and the acknowledged ±1-2% reflectivity uncertainty add to the case for caution.\n\nWho should read it: people working on Ru mono-pnictides or charge-ordering in weakly correlated metals. The LDA comparison and the LT gap data are worth having. I'd send it to a referee, but I'd expect revision on the fitting analysis before the anisotropic onset claim is accepted.\n\nRecommendation: engage with it, but don't take the onset temperatures at face value.","headline":"Solid LDA-backed optical study of RuAs, but the headline anisotropic gap-opening temperatures rely on an unvalidated Drude-Lorentz fit.","tokens_in":10432,"tokens_out":3991,"would_cite":true,"duration_ms":42174,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["71.30.+h","78.20.Ci"],"model":"deepseek-v4-flash","headline":"The central claim is that RuAs's two-step metal-insulator transition is driven by anisotropic gap opening: the c-axis gap appears near 250 K, while the b-axis gap appears only near 200 K.","keywords":["RuAs","metal-insulator transition","anisotropic energy gap","two-step phase transition","charge ordering","optical conductivity","photoelectron spectroscopy","LDA band structure"],"falsifier":"A single experiment that would settle it: polarized optical conductivity on detwinned crystals with the low-energy reflectance measured directly down to a few meV instead of extrapolated; if the two polarizations show the same gap-onset temperature, or if the c-axis gap opens below 240 K, the anisotropic two-step scenario fails.","tokens_in":9393,"feed_emoji":"🔬","tokens_out":6668,"duration_ms":61671,"temperature":0.7,"pith_summary":"This paper uses polarized optical conductivity and photoelectron spectroscopy to show that in the two-step metal-insulator transition of RuAs, the energy gap along the c axis starts opening near the higher transition temperature (~250 K) while the gap along the b axis appears only below the lower transition temperature (~200 K). The authors argue that this directional gap anisotropy is the origin of the two-step phase transition, driven by two separate charge-ordering events with different ordering directions. They also find that both the high-temperature and low-temperature electronic structures, including the gap sizes, are reproduced by simple LDA band calculations, so strong electron correlations are not needed. The middle phase remains structurally uncharacterized, but the optical data reveal that it is where the c-axis gap begins to open.","feed_headline":"In RuAs, the c-axis gap opens 50 K before the b-axis gap","feed_subtitle":"Polarized optics show the two-step transition in RuAs comes from gap opening at different temperatures.","key_machinery":"The central tool is temperature-dependent polarized optical conductivity σ(ω) obtained from Kramers-Kronig analysis of reflectivity, fitted with one Drude and two or three Lorentz oscillators. The load-bearing quantities are the energy of the lowest-energy Lorentz peak and the center of gravity of the spectrum below 1.2 eV, tracked versus temperature and fitted to the BCS gap function to assign each axis its own gap-opening temperature.","core_discovery":"The central discovery is that the metal-insulator transition in RuAs is not a single isotropic event: the polarized optical conductivity shows that the charge gap opens at about 240 ± 10 K along the c axis but only below about 205 K along the b axis, approximately matching the two structural transition temperatures T_MI1 (~250 K) and T_MI2 (~200 K). The lowest Lorentz oscillator peak and the spectral center of gravity track a BCS-like gap function with these two opening temperatures, and the resulting 2Δ/k_B T_c of about 19 rules out a simple Peierls picture. The authors attribute the two-step transition to successive charge orderings along the two directions, consistent with the 3×3×3 superlattice that contains different charge periodicities along c and b.","pith_inferences":["A natural testable extension is to perform the same polarized optical conductivity measurement on detwinned crystals of RuP; if the gap-opening temperatures do not show a similar directional hierarchy, the two-step transitions in RuAs and RuP may have different microscopic origins.","The proposed scenario implies that only the c-axis charge order exists in the incommensurate middle phase, so resonant X-ray scattering at the Ru L-edge could look for that partial ordering and its incommensurate modulation.","Because the gap-opening temperatures are read through a BCS-type mean-field fit, a simultaneous resistivity and optics measurement on the same sample could check whether the extracted onset temperatures coincide with the thermodynamic first-order transitions."],"forward_implications":["If the claim is right, the two-step MIT in RuAs is a sequence of two direction-specific charge-ordering transitions, with the c-axis ordering setting in at T_MI1 and the b-axis ordering at T_MI2.","The success of LDA in reproducing the high- and low-temperature optical and photoemission spectra indicates that the gap formation is a band-structure effect, such as superlattice folding, rather than a correlation-driven Mott-like mechanism.","The large 2Δ/k_B T_c ≈ 19 ratio shows the gap is not purely Peierls-like, so charge ordering or charge pairing must contribute, and the same logic should be tested in the isostructural compound RuP.","Temperature-dependent Drude weight and photoelectron intensity near the Fermi level both decrease through the transition, directly linking the gap opening to the observed rise in electrical resistivity."],"supporting_citations":[{"why":"supplies the high- and low-temperature crystal structures, the transition temperatures T_MI1 and T_MI2, and the 3×3×3 superlattice charge-ordering pattern onto which the gap anisotropy is mapped.","marker":"[10]"},{"why":"provides the LDA-based polarized optical conductivity for the high-temperature phase used as the baseline comparison.","marker":"[11]"},{"why":"gives the Kramers-Kronig analysis method used to convert reflectivity spectra into optical conductivity.","marker":"[12]"},{"why":"supplies the Drude-Lorentz oscillator formula used to fit the optical conductivity and extract the lowest Lorentz peak energy.","marker":"[21]"},{"why":"establishes the BCS-like gap function for charge-density-wave and spin-density-wave systems that the paper uses to determine gap-opening temperatures.","marker":"[22]"},{"why":"demonstrates the same BCS-like gap-function fitting for a charge-ordered compound, supporting the procedure applied to RuAs.","marker":"[23]"}],"fun_headline_variants":["RuAs gap opens along c before b: two-step MIT","Two-step RuAs MIT: anisotropic gap opening via polarized optics","RuAs MIT: gap opens at 240 K along c, 205 K along b","c-axis gap opens ~35 K before b-axis in RuAs MIT"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The central assumption is that the fitted lowest Lorentz peak energy and the spectrum's center of gravity are faithful measures of the true energy gap, so their temperature-dependent rise marks when each gap actually opens.","fun_headline_variants_meta":{"raw":{"variants":["RuAs gap opens along c before b: two-step MIT","Two-step RuAs MIT: anisotropic gap opening via polarized optics","RuAs MIT: gap opens at 240 K along c, 205 K along b","c-axis gap opens ~35 K before b-axis in RuAs MIT"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000673,"raw_usage":{"total_tokens":3050,"prompt_tokens":918,"completion_tokens":2132,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":534,"completion_tokens_details":{"reasoning_tokens":2053}},"tokens_in":534,"tokens_out":2132,"duration_ms":14154,"temperature":1.0,"reasoning_tokens":2053,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:17:02.535063+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A single experiment that would settle it: polarized optical conductivity on detwinned crystals with the low-energy reflectance measured directly down to a few meV instead of extrapolated; if the two polarizations show the same gap-onset temperature, or if the c-axis gap opens below 240 K, the anisotropic two-step scenario fails.","supporting_citations":[{"cited_title":"Kotegawa, K","cited_arxiv_id":null,"evidence_quote":"supplies the high- and low-temperature crystal structures, the transition temperatures T_MI1 and T_MI2, and the 3×3×3 superlattice charge-ordering pattern onto which the gap anisotropy is mapped."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"provides the LDA-based polarized optical conductivity for the high-temperature phase used as the baseline comparison."},{"cited_title":"Wooten, Optical Properties of Solids (Academic Press, New York, 1972)","cited_arxiv_id":null,"evidence_quote":"supplies the Drude-Lorentz oscillator formula used to fit the optical conductivity and extract the lowest Lorentz peak energy."},{"cited_title":"Dressel, L","cited_arxiv_id":null,"evidence_quote":"establishes the BCS-like gap function for charge-density-wave and spin-density-wave systems that the paper uses to determine gap-opening temperatures."},{"cited_title":"Kimura, T","cited_arxiv_id":null,"evidence_quote":"demonstrates the same BCS-like gap-function fitting for a charge-ordered compound, supporting the procedure applied to RuAs."}],"review_version":1}