{"id":"3954e7ed-8188-4162-af01-dfec767ead2b","arxiv_id":"2501.08705","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Sodium intercalation shifts the Reststrahlen band of α-V2O5 by about 30 cm-1 while preserving picosecond phonon polariton lifetimes.","lead":"This paper shows that inserting sodium atoms between the layers of a van der Waals crystal shifts its mid-infrared response band by about 30 wavenumbers, while the crystal still guides phonon polaritons with a lifetime of about 4 picoseconds. A smart generalist would read it because it suggests a practical way to tune infrared nanophotonics materials without sacrificing their low optical losses.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lifetime comparison is the weak link: the 4 vs 6 ps values come from flakes of different thickness measured at different frequencies, and the conversion uses a damping parameter tuned to the lifetime.","rationale":"The manuscript has two distinct claims: a broad RB shift and preserved ultra-low-loss PhPs. The RB shift is supported by nanoFTIR line scans (Fig. 3b,d), measured dispersions (Fig. 4a), and ab initio permittivities (Fig. 4b); although the magnitudes differ (~30 vs ~50 cm^-1), the direction and order of magnitude agree, so that claim is credible. The lifetime claim is much less secure. The 6±1 vs 4±1 ps values are not directly measured lifetimes but are derived as L/v_g from fringe fits. The comparison changes two variables at once (thickness 130 vs 107 nm; frequency 1010 vs 973 cm^-1) in a configuration where both are known to affect the loss. Further, the dielectric damping γ_z was explicitly adjusted to 1.5 cm^-1 to reproduce the lifetime, so using that model to convert fringe decay to lifetime risks circular reasoning. These issues do not undermine the spectral-tuning result, but they do mean the central 'ultra-low losses' claim should be conditional on a matched-control measurement. The reader's conditional verdict is therefore appropriate; no change is needed.","tokens_in":11130,"tokens_out":7435,"duration_ms":80879,"concrete_test":"Re-measure PhP lifetimes in pristine and Na-intercalated flakes of matched thickness (e.g., 110–120 nm) at frequencies with the same normalized detuning from the LO phonon inside RB1 and RB'1, and extract τ using v_g from the measured dispersion only, without setting γ_z=1.5 cm^-1. If the two lifetimes remain within 1 ps, the confound is benign; if they separate by more than 2 ps or the pristine lifetime becomes much larger, the paper's similarity claim should be revised.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Intercalation preserving ultra-low losses is supported only by the 4±1 ps vs 6±1 ps comparison in Fig. 5. The two measurements are not like-for-like: the pristine α-V2O5 flake is 130 nm thick and measured at ω0=1010 cm^-1, while the α'-(Na)V2O5 flake is 107 nm and measured at 973 cm^-1. For slab phonon polaritons, the fringe decay length and the τ=L/v_g conversion both depend on slab thickness (confinement and substrate losses) and on frequency through v_g(ω). The 2 ps difference and overlapping error bars therefore do not isolate the effect of Na intercalation; a larger intercalation-induced loss change could be hidden by the thickness/frequency offsets. In addition, the Methods state that γ_z=1.5 cm^-1 was adopted 'to better fit the experimental lifetime values'; if the group velocity used in τ=L/v_g is obtained from that adjusted dielectric model, the reported lifetimes are not an independent measurement of loss. The RB-shift part of the paper is less affected, but the 'ultra-low losses preserved' claim is not established by the data as presented.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports that Na intercalation of the van der Waals semiconductor α-V2O5 (forming α'-(Na)V2O5) shifts the mid-infrared Reststrahlen band RB1 to lower frequencies by about 30 cm^-1 (roughly 60% of the initial RB width), as directly observed by nanoFTIR line scans and s-SNOM dispersion measurements. The authors also report that phonon polaritons in the intercalated crystal remain ultra-low-loss, with a lifetime of 4 ± 1 ps inferred from antenna-launched fringe decay, compared with 6 ± 1 ps in pristine α-V2O5. Ab initio phonon calculations reproduce the direction of the RB shift and are used to explain the disappearance of RB2 and RB3 in the intercalated material.","tokens_in":11348,"tokens_out":3180,"duration_ms":32541,"significance":"If the central claims hold, the paper demonstrates a broadly useful tuning mechanism for phonon polaritons in van der Waals crystals, addressing a recognized limitation of these materials. The strengths of the paper are the direct near-field observation of the RB shift, the use of nanoFTIR to map the spectral position of the bands, the mapping of in-plane anisotropic propagation, and the ab initio calculations that corroborate the shift direction. However, the 'ultra-low losses preserved' claim is less strongly supported: it rests on a single lifetime comparison between flakes of different thickness measured at different frequencies, and the lifetime extraction uses a damping parameter that is itself tuned to reproduce the experimental lifetimes. The RB-shift result is more robust, but the quantitative comparison between the measured (~30 cm^-1) and calculated (~50 cm^-1) shifts needs tightening.","major_comments":[{"comment":"The claim that intercalation preserves ultra-low losses rests on a single comparison between a 130-nm-thick α-V2O5 flake measured at 1010 cm^-1 and a 107-nm-thick α'-(Na)V2O5 flake measured at 973 cm^-1. For slab phonon polaritons, the fringe decay length and the conversion τ = L/v_g depend on slab thickness (through confinement and substrate losses) and on frequency through v_g(ω). The 2 ps difference with overlapping error bars therefore does not isolate the effect of Na intercalation; a larger intercalation-induced loss increase could be hidden by the thickness and frequency offsets. Please provide like-for-like measurements (same thickness and frequency, or a quantitative model that corrects for these differences) before claiming that intercalation preserves ultra-low losses.","section":"Fig. 5 and 'Finally, we study and compare...' paragraph"},{"comment":"The lifetime extraction is partly circular. The Methods state that γ_z = 1.5 cm^-1 is adopted 'to better fit the experimental lifetime values', and the group velocities used in τ = L/v_g are obtained from the same Lorentz-oscillator permittivity. Thus the reported 4 ± 1 ps and 6 ± 1 ps lifetimes are not an independent measurement of loss. Please compute the group velocity from the unadjusted permittivity (γ_z = 2.0 cm^-1) or directly from the measured dispersion, and report whether the conclusion changes.","section":"Methods, 'Dielectric function of α-V2O5'"},{"comment":"The ab initio calculation gives a center-to-center RB'1 shift of about 50 cm^-1, while the measured shift is about 30 cm^-1. The text calls this 'good agreement', but no uncertainty is given for the measured RB edges and the 20 cm^-1 discrepancy is not discussed. Since the tunability claim is quantified by this shift, please provide a quantitative comparison with error bars and a brief discussion of possible sources of the discrepancy (e.g., Na concentration, anharmonicity, or approximations in the Lorentz model).","section":"Fig. 4 and 'To better understand the polaritonic effects...' paragraph"},{"comment":"The absence of RB'2 and RB'3 in the intercalated crystal is explained after the fact by the calculated narrowing and spectral shifts of these bands. Because the same experimental data are used to infer both the shift of RB'1 and the absence of the other bands, an independent check is needed to rule out that the missing bands are due to the specific flake thickness, the limited spectral window, or an inhomogeneous intercalation. A wider-range nanoFTIR scan or measurements on a second flake of different thickness would strengthen this point.","section":"Fig. 3 and 'We note that the nanoFTIR image in Fig. 3b...' paragraph"}],"minor_comments":[{"comment":"There is a figure-callout inconsistency: the text refers to 'Figure 3b' for the nanoFTIR scans of the intercalated flake, but in the figure caption panel b is the pristine α-V2O5 scan and panel d is the α'-(Na)V2O5 scan. Please correct the callouts.","section":"Fig. 3 caption and main text"},{"comment":"The displayed Drude-Lorentz equation is poorly typeset, with superscripts and subscripts garbled. Please ensure the equation is rendered with clear indices for the three axes.","section":"Methods, 'Dielectric function of α-V2O5'"},{"comment":"The phrases 'ultra-low losses' and 'ultra-long lifetime' are used for lifetimes of 4–6 ps that are inferred through a model with a tuned damping parameter; consider softening the wording or adding a caveat that the lifetimes are model-derived.","section":"Abstract and introduction"},{"comment":"In the lifetime-comparison paragraph, the two flakes are described as having 'similar thicknesses of 130 and 107 nm'. A 23 nm difference is not negligible for slab polaritons; please either use the same-thickness data or justify the approximation quantitatively.","section":"Full text, 'similar thicknesses'"}],"recommendation":"major_revision","confidential_remarks":"The RB-shift demonstration is the strongest part of the paper and is likely suitable for the journal after a focused revision. The main risk is the lifetime-preservation claim, which rests on a non-like-for-like comparison and a partially circular extraction. I would support acceptance if the authors provide a proper like-for-like loss measurement or a quantitative correction, and if they address the γ_z tuning issue explicitly."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"What you should know: this paper demonstrates a genuinely new way to tune phonon polaritons. Sodium intercalation in α-V2O5 red-shifts the Reststrahlen band by about 30 cm-1 (roughly 60% of the original band width), and the intercalated crystal still supports in-plane anisotropic phonon polaritons with lifetimes in the picosecond range. That combination—broad spectral shift with preserved low-loss propagation—is the new result, and it is supported by direct near-field imaging, not inferred from a fit. The authors also add α-V2O5 itself as a new polaritonic van der Waals crystal, with elliptic and hyperbolic bands mapped out along different crystal axes.\n\nThe strongest evidence is the nanoFTIR line scans in Fig. 3 and the s-SNOM dispersions in Fig. 4: the shift is seen directly in the spectral position of the fringes, and the ab initio phonon calculation reproduces the direction of the shift, even though it overestimates the magnitude (about 50 cm-1 versus the measured ~30). That near-field observation alone justifies taking the claim seriously.\n\nWhere the paper gets soft is the lifetime comparison. The headline \"ultra-low losses preserved\" rests on 4 ± 1 ps in the intercalated flake versus 6 ± 1 ps in the pristine one. But those two values come from flakes of different thickness (107 nm vs 130 nm) measured at different frequencies (973 vs 1010 cm-1). For slab phonon polaritons, both the fringe decay length and the conversion to lifetime via group velocity depend on thickness and frequency, so the comparison is not like-for-like. More concerning, the Methods state that γ_z = 1.5 cm-1 was adopted \"to better fit the experimental lifetime values.\" If that adjusted damping constant enters the group velocity used to get τ = L/v_g, the lifetime is partly circular. The 2 ps difference with overlapping error bars does not isolate the effect of intercalation.\n\nThe RB-shift claim itself is much less affected by these issues. The shift is directly measured in the near-field spectra, so that part holds up. The main soft spots are the lifetime extraction, the published-vs-calculated shift discrepancy, and the lack of public data or code despite the availability statements.\n\nWho this is for: anyone working on phonon polaritons, tunable infrared nano-optics, or van der Waals materials. The central result is novel and the near-field evidence is credible, so the paper deserves serious peer review. A good referee should ask for a cleaner lifetime analysis—ideally same-thickness flakes or a model that propagates uncertainties through the whole extraction—and for the data/code to be released. I would not block acceptance on the shift, but the lifetime claim needs to be tightened before publication.","headline":"Intercalation shifts the Reststrahlen band of α-V2O5 by ~30 cm-1 while preserving picosecond phonon polariton lifetimes—a real result, but the lifetime comparison is not yet airtight.","tokens_in":12026,"tokens_out":1399,"would_cite":true,"duration_ms":15060,"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":"Intercalating sodium into α-V2O5 shifts its phonon-polariton band by about 30 cm−1 while lifetimes stay in the picosecond range.","keywords":["phonon polaritons","van der Waals crystals","intercalation","Reststrahlen band","alpha-V2O5","s-SNOM","nanoFTIR","mid-infrared nanophotonics"],"falsifier":"Measure the same α'-(Na)V2O5 flakes at several thicknesses and frequencies with a different technique that extracts the complex permittivity directly, such as FTIR ellipsometry, and compute the lifetime from that permittivity; if the resulting lifetime falls well below the picosecond range or differs markedly from pristine α-V2O5, the central claim fails.","tokens_in":10924,"feed_emoji":"🔬","tokens_out":7245,"duration_ms":64607,"temperature":0.7,"pith_summary":"This paper establishes that intercalating sodium atoms into the layered semiconductor α-V2O5 is a workable route to spectrally shift its Reststrahlen bands, and therefore the phonon polaritons inside them, without destroying their low-loss character. If correct, it gives a concrete method for placing phonon-polariton devices at frequencies the parent crystal cannot reach, simply by choosing the intercalant and its concentration. The paper supports the claim with near-field images and nanoFTIR line scans showing the first band moving roughly 30 cm−1 red, with ab initio phonon calculations that reproduce the shift, and lifetime measurements of 4±1 ps in the intercalated crystal versus 6±1 ps in pristine α-V2O5.","feed_headline":"Intercalation shifts phonon-polariton bands by 30 cm-1","feed_subtitle":"Sodium atoms between V2O5 layers retune mid-infrared polaritons while lifetimes stay in the picosecond range.","key_machinery":"The central object is the Na-intercalated van der Waals crystal α'-(Na)V2O5, in which sodium atoms sit between the weakly bound layers and modify the phonon modes and effective charges, shifting the Reststrahlen bands. The experimental machinery is s-SNOM and nanoFTIR polariton interferometry, where tip-launched and antenna-launched polaritons interfere with back-reflected fields; the fringe period gives the polariton wavelength and dispersion, while the fringe decay, fitted with an exponentially damped sine wave corrected by a 1/√x geometric spreading factor, gives the propagation length. The propagation length is converted into a lifetime using the measured group velocity via τ = L/vg.","core_discovery":"The paper demonstrates that intercalating Na atoms between the van der Waals layers of α-V2O5 red-shifts the first Reststrahlen band by about 30 cm−1, roughly 60% of the initial band width, and that the phonon polaritons supported in the resulting α'-(Na)V2O5 crystal remain anisotropic and ultra-low-loss, with a measured lifetime of 4±1 ps compared with 6±1 ps in pristine α-V2O5. It further shows, through first-principles phonon calculations feeding a Lorentz-oscillator permittivity model, that the other two Reststrahlen bands are also modified by intercalation, explaining why they become nearly invisible in the measurements. The central conclusion is that intercalation can be used as a broad spectral tuning mechanism for phonon polaritons while preserving their low-loss propagation.","pith_inferences":["The same intercalation route could be tried on other polar van der Waals crystals, such as α-MoO3, where the variety of intercalant ions and concentrations may tune the Reststrahlen bands; the paper states this as an expectation, not a demonstrated result.","The 30 cm−1 experimental shift versus roughly 50 cm−1 in calculation leaves room for testing how sodium concentration, staging, or strain adjusts the TO–LO splitting, which could refine the model.","Because α-V2O5 can also be electrically doped, combining intercalation with electrostatic gating could turn the static spectral shift into a dynamically switchable polariton frequency; this is an implication beyond the paper's demonstrated static tuning.","A direct time-domain measurement of the polariton decay, rather than fringe-decay fitting, would be the cleanest check of whether the 4±1 ps lifetime is intrinsic or partly an artifact of the extraction model."],"forward_implications":["α-V2O5 supports anisotropic phonon polaritons in three Reststrahlen bands, with elliptic in-plane propagation in RB1 and hyperbolic behavior in RB2 and RB3.","Intercalating Na atoms red-shifts RB1 by roughly 30 cm−1, moving the polariton dispersion to lower frequencies, with ab initio calculations giving a similar larger shift of about 50 cm−1.","Phonon polaritons in α'-(Na)V2O5 remain anisotropic and low-loss, with a measured lifetime of 4±1 ps and quality factor Q≈2.5, close to the pristine values of 6±1 ps and Q≈3.5.","If the method generalizes to other ions and layered hosts, the spectral position of phonon polaritons could be chosen by intercalation chemistry rather than by choosing a different material."],"supporting_citations":[{"why":"Supplies the method and benchmark for imaging in-plane anisotropic ultra-low-loss phonon polaritons in a van der Waals crystal.","marker":"[2]"},{"why":"Establishes isotopically pure h-BN as an ultra-low-loss polariton reference, supporting the isotopic-purity argument for V2O5.","marker":"[3]"},{"why":"Previous demonstration of tunable hyperbolic phonon polaritons in transition metal oxides, the prior intercalation-shift result this work extends.","marker":"[4]"},{"why":"Framework for mid-infrared biaxial hyperbolic van der Waals crystals used to interpret the anisotropic propagation in α-V2O5.","marker":"[5]"},{"why":"Source for the anisotropy and exfoliation behavior of thin α-V2O5 flakes used as samples.","marker":"[14]"},{"why":"Provides the TO/LO frequencies and damping constants for the Lorentz-oscillator permittivity of α-V2O5.","marker":"[15]"},{"why":"Supplies the damped-sine-wave fitting with the 1/√x spreading correction used to extract propagation lengths and lifetimes.","marker":"[28]"},{"why":"Crystal-growth method for NaV2O5 used to make the intercalated α'-(Na)V2O5 samples.","marker":"[32]"},{"why":"The analogy justifying the reduced damping γz=1.5 cm−1 used to match measured lifetimes.","marker":"[33]"}],"fun_headline_variants":["Sodium intercalation retunes polariton bands by 30 cm-1","Intercalation shifts phonon polaritons by 30 cm-1, keeps low loss","Sodium intercalation broadens polariton tuning without loss","Intercalation retunes polariton bands by 30 cm-1, preserving low loss","Intercalation widens mid-IR polariton tunability, keeps low loss"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The central claim leans on converting near-field fringe decay into a lifetime through a chosen damping parameter and a damped-sine model; if that conversion is not right, the 4±1 ps value and the 'intercalation preserves low losses' conclusion do not follow.","fun_headline_variants_meta":{"raw":{"variants":["Sodium intercalation retunes polariton bands by 30 cm-1","Intercalation shifts phonon polaritons by 30 cm-1, keeps low loss","Sodium intercalation broadens polariton tuning without loss","Intercalation retunes polariton bands by 30 cm-1, preserving low loss","Intercalation widens mid-IR polariton tunability, keeps low loss"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000651,"raw_usage":{"total_tokens":2981,"prompt_tokens":938,"completion_tokens":2043,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":554,"completion_tokens_details":{"reasoning_tokens":1937}},"tokens_in":554,"tokens_out":2043,"duration_ms":16320,"temperature":1.0,"reasoning_tokens":1937,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T20:19:40.182757+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the same α'-(Na)V2O5 flakes at several thicknesses and frequencies with a different technique that extracts the complex permittivity directly, such as FTIR ellipsometry, and compute the lifetime from that permittivity; if the resulting lifetime falls well below the picosecond range or differs markedly from pristine α-V2O5, the central claim fails.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes isotopically pure h-BN as an ultra-low-loss polariton reference, supporting the isotopic-purity argument for V2O5."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Previous demonstration of tunable hyperbolic phonon polaritons in transition metal oxides, the prior intercalation-shift result this work extends."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Framework for mid-infrared biaxial hyperbolic van der Waals crystals used to interpret the anisotropic propagation in α-V2O5."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Source for the anisotropy and exfoliation behavior of thin α-V2O5 flakes used as samples."},{"cited_title":"& Vennik, J","cited_arxiv_id":null,"evidence_quote":"Provides the TO/LO frequencies and damping constants for the Lorentz-oscillator permittivity of α-V2O5."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the damped-sine-wave fitting with the 1/√x spreading correction used to extract propagation lengths and lifetimes."}],"review_version":1}