{"id":"b4c87106-73ac-491b-8744-e8df1f0f5f7d","arxiv_id":"2507.20896","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":4,"one_line_summary":"Isotopic substitution of 18O for 16O in beta-Ga2O3 redshifts hyperbolic shear polaritons by about 40 cm^-1 while preserving their directional propagation.","lead":"This paper shows that replacing oxygen-16 with oxygen-18 in beta-gallium oxide shifts the frequency range of hyperbolic shear polaritons, highly directional nanoscale infrared light waves, by about 40 inverse centimeters without changing their propagation pattern. The result adds isotopic substitution as a practical tuning knob for infrared nanophotonics and offers a near-field imaging shortcut for detecting such frequency shifts.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Semi-infinite-film assumption for the 1.2 µm 18O epilayer depends on an inconsistent skin-depth estimate; substrate leakage could mix 16O bands into the 18O near-field data.","rationale":"The reader identified exactly this assumption as the weakest point, and the main text supports that framing: the semi-infinite claim is asserted with an approximate skin-depth estimate (SI S9), and the manuscript itself acknowledges (SI S3) that larger discs produce patterns with a film/substrate coupled component. The central claim is nevertheless strongly supported by independent evidence: FT-IR fits give TO shifts of ~5% for the high-frequency modes, DFT reproduces those shifts, and the near-field angle data match the FT-IR-derived γ(ω) and α(ω) curves. The residual concern is not that the film is certainly not semi-infinite, but that the quantitative margin for the 1.2 µm film at the specific frequencies and antenna size used is not demonstrated, and the inconsistency in SI S9 between the 0.23 µm, 0.46 µm, and 0.84 µm numbers (with a 4 µm antenna, whereas the experiment used 2 µm) leaves the margin uncertain. A concrete layered transfer-matrix validation would settle this cleanly. I keep the reader's CONDITIONAL verdict: the concern is real but likely addressable, and the agreement among three methods is strong enough that REJECT is not warranted; if the stack simulation resolves the ambiguity, ACCEPT would be justified.","tokens_in":30523,"tokens_out":2156,"duration_ms":21882,"concrete_test":"Perform a transfer-matrix calculation of the launched near-field response for the actual sample stack (air / 2 µm Au disc / 1.2 µm 18O bGO / 16O bGO substrate) at the measured FEL frequencies, using the fitted permittivities in Table S1, and compute the spatial intensity pattern and its angular dips. Compare the extracted ray angles with the angles predicted for a semi-infinite 18O crystal. If the stack simulation reproduces the semi-infinite predictions to within the experimental error bars at all measured frequencies, the semi-infinite-film assumption is validated; if the stack pattern shows detectable 16O or film/substrate contributions at any measured frequency, the central claim requires re-analysis.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim — that the 18O near-field images show HShPs of 18O bGO redshifted by ~40 cm−1 — rests on the claim in Section 2 that a 1.2 µm thick 18O bGO film acts as a semi-infinite crystal. This relies on the estimate in SI S9 that Au discs excite momentum ~10 k0, giving a field skin depth of about 0.5 µm into the film. However, SI S9 contains an internal inconsistency: it first quotes a penetration depth dbGO = 0.23 µm for |εzz| ≈ 4, then says the skin depth (fields) is 0.46 µm, then states that for a 4 µm-thick Au disc the skin depth into bGO is 0.84 µm assuming kx,0 ≈ 5 — despite the main text using 2 µm discs at ~10 k0. The factor-of-two distinction between intensity and field decay, the antenna-size dependence, and the value of kx,0 are not tied to a quantitative near-field calculation for polaritons in a layered 18O/16O structure. Since the 16O substrate near 682 cm−1 is close to its own reststrahlen features, any substrate leakage could distort the observed ray directions and angles. If at some measured frequencies part of the signal is a 16O substrate mode or a coupled film/substrate mode, the extracted γ(ω) and α(ω) would no longer be purely 18O quantities, and the model-free shift estimate could be biased.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports spectral tuning of hyperbolic shear polaritons (HShPs) in monoclinic beta-Ga2O3 by substituting 18O for 16O. Near-field imaging of a 1.2 micrometer 18O epilayer homoepitaxially grown on a 16O substrate shows ray-like polariton patterns at 667-706 cm^-1 that closely resemble 16O patterns shifted by about 40 cm^-1, with the 16O image at 720 cm^-1 matching the 18O image at 682 cm^-1. Polarized FT-IR reflectance and DFT calculations yield TO phonon frequency redshifts of roughly 5% for the high-frequency Bu modes, while the optical-axis dispersion and hyperbola opening angle remain essentially unchanged. The authors argue that the near-field images alone provide a model-free estimate of the isotope-induced frequency shift, without requiring knowledge of the 18O dielectric tensor.","tokens_in":31068,"tokens_out":10511,"duration_ms":122516,"significance":"If the near-field interpretation is correct, the work demonstrates a practical route to spectral tuning of highly directional shear polaritons in a non-van-der-Waals, low-symmetry 3D crystal, with potential applications in nanophotonic devices operating in previously inaccessible frequency ranges. The study is strengthened by the combination of three independent approaches: real-space near-field imaging, azimuth-dependent FT-IR reflectance fitted with a multilayer model, and DFT calculations with isotope masses changed from first principles. The claim that the optical-axis orientation and hyperbola opening angle are nearly unchanged under isotopic substitution is supported by both the FT-IR-derived permittivity and the DFT results. The near-field analysis is genuinely model-free in the sense that the ray directions and the resulting gamma and alpha values are extracted without fitting a dielectric tensor to the near-field data, which is a useful methodological contribution for thin epitaxial layers and small samples.","major_comments":[{"comment":"The claim that the 1.2 micrometer 18O epilayer acts as a semi-infinite crystal is not supported by a consistent quantitative estimate. The main text states that 2 micrometer Au discs launch momenta around 10 k0 with a skin depth of about 0.5 micrometers into the film, while SI S9 defines d = 1/kappa with kappa = k0 sqrt(kx0^2 - epsilon), which for k0 = 0.44 micrometer^-1, kx0 = 10, and epsilon_zz about -4 gives d about 0.22 micrometer; the text then quotes 0.46 micrometer as a factor-of-two 'skin depth' and also gives 0.84 micrometer for a '4 micrometer thick' disc at kx0 about 5. Because the assertion that the 16O substrate does not contribute to the 18O near-field images is load-bearing for the model-free shift estimate, please replace this estimate with a quantitative layered calculation, for example a transfer-matrix or full-wave simulation of the disk-launched field at the measured frequencies, and state unambiguously whether field-amplitude or intensity decay is being quoted.","section":"Section 2; SI S9"},{"comment":"The signal processing is not identical for the two isotopes: the 18O images in Fig. 2 are analyzed as the raw O2A amplitude, whereas the 16O images in Fig. S4.1 are processed as S2 = -O2A * cos(O2P) to compensate for self-homodyne mixing of amplitude and phase. Since the central comparison of gamma and alpha in Fig. 4 mixes these two processing routes, and the Methods section states that amplitude and phase cannot be separated in the self-homodyne scheme, the relative extracted angles could be biased. Please apply the same processing to both isotopes, or demonstrate explicitly (for example by re-extracting the ray angles from both channels for at least one frequency per isotope) that the phase correction does not change the extracted directions.","section":"Methods; SI S1 and SI S4"},{"comment":"The FT-IR fit for 18O is initialized from DFT TO frequencies, and Table 1 is then presented as an FT-IR/DFT confirmation of the isotope shift. Please state explicitly which parameters were free in the final fit and provide a sensitivity test, such as refitting the 18O reflectance starting from the 16O TO frequencies, to show that the fitted 18O TO positions are determined by the reflectance data rather than by the DFT seed. This would also clarify how independent the far-field confirmation is from the ab initio prediction.","section":"SI S5; Table 1"}],"minor_comments":[{"comment":"The text says the calculations used 'a X x X x X k-grid', which appears to be a placeholder; please supply the actual k-grid and the supercell sizes used for the finite-displacement phonon and anharmonic calculations.","section":"Methods, DFT calculations"},{"comment":"There are typos in the sentence about the hyperbolic range: 'spannig' should be 'spanning' and 'bonunded' should be 'bounded'.","section":"Section 4"},{"comment":"The caption says 'blue and orange lines, respectively' while the text says 'orange and blue curves'; please make the color-to-isotope assignment consistent in both places.","section":"Fig. 3 caption"},{"comment":"The text refers to 'the principal eigenvector of Re(epsilon_xy)', but the optical-axis direction should be determined from the full in-plane real permittivity tensor, not from the xy component alone; please correct or clarify this notation.","section":"Eq. (1) and surrounding text"},{"comment":"In addition to the consistency issue raised above, 'a 4 micrometer thick Au disc' should presumably read 'a 4 micrometer diameter Au disc', and the factor-of-two distinction between intensity and field decay should be defined explicitly or removed.","section":"SI S9"},{"comment":"Reference 13 appears incomplete, as it lists authors and an arXiv identifier but no title; please check whether this citation is appropriate for the claim it supports.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The paper is within the scope of the journal and the result is timely and plausible. The central claim of a roughly 40 cm^-1 isotope-induced shift is supported by independent FT-IR and DFT evidence, so I do not see a need for rejection. The main revision should focus on the near-field methodology: a quantitative layered calculation for the film-substrate system and a consistent treatment of the self-homodyne signal. If those are provided, I expect the paper to be suitable for publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nBottom line: this paper convincingly shows a ~40 cm^-1 redshift of hyperbolic shear polaritons in 18O bGO, and I think the central claim holds. The near-field images at 682 and 720 cm^-1, the frequency-dependent ray angles, the FT-IR oscillator fits, and DFT all point to the same shift. DFT just changes the oxygen masses and predicts the shift; no fitted parameter produces the 40 cm^-1 value. That is real evidence.\n\nWhat is new: first isotope tuning of shear polaritons in a 3D monoclinic crystal, and a neat model-free way to estimate the spectral shift from the ray angles alone, without knowing the 18O dielectric tensor. The paper is honest about the self-homodyne limitation and about the FT-IR cutoff issues for modes #7 and #8.\n\nSoft spots, in proportion. The SI S9 skin-depth estimate is internally inconsistent: it quotes d = 0.23 µm for intensity, then 0.46 µm skin depth for fields, then 0.84 µm for a 4 µm Au disc at kx,0 ~ 5, while the main text uses 2 µm discs at ~10 k0 and claims ~0.5 µm. That is sloppy and should be cleaned up. But I do not think it sinks the paper: at 682 cm^-1 the 16O substrate is not in its hyperbolic band (16O HShPs appear around 705-725 cm^-1), so substrate leakage would not produce a competing set of rays at the 18O measurement frequencies. The assumption is plausible even if the supporting estimate is messy.\n\nThe self-homodyne processing also has an unexplained isotope asymmetry: SI S4 says 16O images were corrected with cos(O2P), while the 18O analysis appears to use raw O2A dips. That deserves an explanation, but the 18O ray patterns are clear enough that I do not see it overturning the result.\n\nThe FT-IR fit seeds 18O TO frequencies from DFT; that is a mild circularity for the far-field model, but the near-field shift is direct and does not depend on the fit. So the load-bearing claim is safe.\n\nWho is this for: phonon-polariton and nanophotonics researchers, plus anyone working on isotope engineering of infrared optical properties. It is a solid, useful contribution to a niche field. I would send it to peer review; the SI inconsistencies and the signal-processing asymmetry should be addressed in revision, but they are fixable. I would cite it if I worked on bGO polaritons.\n\nRecommendation: engage, accept with minor revision.","headline":"A convincing demonstration that isotope substitution shifts hyperbolic shear polaritons in beta-Ga2O3 by ~40 cm^-1, with three independent methods agreeing; the main soft spot is a sloppy skin-depth estimate in the SI, not the central claim.","tokens_in":31567,"tokens_out":2626,"would_cite":true,"duration_ms":29654,"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":"Substituting 18O for 16O in beta-Ga2O3 redshifts its hyperbolic shear polariton band by about 40 cm^-1 while leaving the polariton geometry essentially unchanged.","keywords":["hyperbolic shear polaritons","beta-Ga2O3","isotopic substitution","near-field optical microscopy","phonon polaritons","reststrahlen band","s-SNOM","spectral tuning"],"falsifier":"Thin the 18O epilayer to about half a micrometer, or excite the same film with a larger antenna that launches lower momenta, and check whether the apparent 40 cm$^{-1}$ shift splits into two bands or moves toward the 16O substrate response; the semi-infinite assumption predicts no substrate contribution, while a substrate-coupling alternative predicts a mixed or thickness-dependent shift.","tokens_in":30375,"feed_emoji":"🔬","tokens_out":13480,"duration_ms":127585,"temperature":0.7,"pith_summary":"This paper reports that substituting the oxygen isotope 18O for natural 16O in the monoclinic crystal $\\beta$-Ga2O3 shifts the hyperbolic shear polariton band downward by roughly 40 cm$^{-1}$, about 5 percent of the frequency of the high-energy transverse-optical phonons that define the band. The shift is practically relevant because hyperbolic shear polaritons, which guide infrared light along nanoscale directional rays, exist only inside narrow reststrahlen bands fixed by the crystal's phonon frequencies; isotopic substitution moves those bands to frequencies the natural crystal cannot access. The authors show the shift can be measured directly from near-field images of the polariton rays, through the orientation of the optical axis and the opening angle of the hyperbolic wave fronts, without fitting an 18O dielectric tensor. Polarized infrared reflectance and density-functional-theory calculations agree with the near-field estimate.","feed_headline":"Oxygen isotope swap tunes polariton band by 40 cm-1","feed_subtitle":"With 18O, beta-gallium oxide guides the same directional infrared modes about 40 cm-1 lower, into frequency range natural bGO cannot reach.","key_machinery":"The central object is the hyperbolic shear polariton (HShP), a surface-bound infrared mode in monoclinic crystals whose in-plane isofrequency contour is a hyperbola with two asymmetric arms; the paper images these modes in real space after launching them from a 2-µm gold disc that acts as a high-momentum antenna. The argument is carried by two angle quantities extracted from the near-field images: the optical-axis dispersion angle $\\gamma(\\omega)$, defined from the eigenvector of $\\mathrm{Re}(\\varepsilon)$ as $\\gamma(\\omega) = \\mathrm{arctan2}(m_x, m_y)$, and the hyperbola opening angle $\\alpha(\\omega) = 2\\arctan\\sqrt{-\\mathrm{Re}(\\varepsilon_{nn})/\\mathrm{Re}(\\varepsilon_{mm})}$. Isotopic substitution enters through the Lorentz-oscillator model of the permittivity, in which the transverse-optical frequencies shift with oxygen mass while the oscillator orientation vectors $S_i$ remain essentially unchanged, so the $\\gamma$ and $\\alpha$ values read from the ray directions already contain the isotope shift without any need to know the 18O dielectric tensor.","core_discovery":"On the paper's own terms, the central discovery is that 18O bGO supports the same hyperbolic shear polaritons as 16O bGO, only shifted about 40 cm$^{-1}$ lower in frequency: a near-field image at 682 cm$^{-1}$ on the 18O epilayer reproduces the pattern seen at 720 cm$^{-1}$ on 16O bGO. The frequency-dependent optical-axis angle and hyperbola half-opening angle extracted from real-space ray propagation match the curves computed from the FT-IR permittivity of 18O bGO, and both are the 16O curves displaced by the same amount. Table 1 quantifies the shift in the in-plane transverse-optical phonons: the high-frequency B$_u$ modes #1-#4 shift by 4.4-5.6% (FT-IR) and 4.9-5.5% (DFT), while the oscillator orientation vectors stay essentially fixed. Because modes #1 and #2 dictate the permittivity in the hyperbolic band, the ~5% shift of these modes yields the observed mid-band shift of about 40 cm$^{-1}$. The authors conclude that isotopic substitution tunes the spectral position of hyperbolic shear polaritons while preserving their propagation characteristics.","pith_inferences":["If the semi-infinite assumption is tested by varying the epilayer thickness, the apparent shift should weaken below roughly half a micrometer; a persistent shift at smaller thickness would instead indicate the isotope effect is robust to substrate coupling.","The same near-field angle-extraction recipe could be applied to other monoclinic or triclinic polar crystals, turning real-space ray angles into a model-free probe of compositional or isotopic changes where the dielectric tensor is unknown.","Isotopically patterned heterostructures, with 16O and 18O regions on one chip, could confine the same type of mode to frequency-selective domains; the paper does not attempt such patterning.","A direct propagation-length measurement in the 18O film would clarify whether the larger fitted damping found for some low-frequency modes alters losses in the shifted band, since the near-field images here are not primarily loss measurements."],"forward_implications":["The 40 cm-1 redshift places the 18O bGO hyperbolic band at frequencies where natural bGO is elliptical, effectively widening the spectral coverage of bGO-based nanophotonics.","Because the optical-axis dispersion and hyperbola opening angle are unchanged, the ray-like propagation design rules derived for natural bGO carry over to the shifted band.","The near-field extraction works on a 1.2 µm epitaxial film, where conventional far-field reflectance fits must model both film and substrate, making it a practical tool for thin-film isotope assessment.","The agreement between FT-IR and DFT relative shifts for the high-frequency modes supports using phonon masses to predict the tuning range before growth."],"supporting_citations":[{"why":"It introduced hyperbolic shear polaritons in low-symmetry crystals and supplies the formalism connecting the permittivity eigenvector to the optical-axis angle.","marker":"23"},{"why":"It established the gold-disc launching scheme and real-space analysis of asymmetric hyperbolic shear polariton propagation in beta-Ga2O3 that the near-field measurements build on.","marker":"25"},{"why":"It reported the molecular-beam epitaxy of 18O bGO and the roughly 5 percent redshift of high-frequency Raman phonons that motivates the expected polariton shift.","marker":"50"},{"why":"It provides the analytical relation between the in-plane permittivity components and the hyperbola opening angle used to compare near-field data with theory.","marker":"61"},{"why":"It supplies the multi-oscillator model and reference infrared parameters for 16O bGO used to fit the polarized FT-IR reflectance.","marker":"62"},{"why":"It reviews the general theory of isotope-induced phonon frequency shifts that underlies the mass-dependent tuning demonstrated here.","marker":"42"},{"why":"It demonstrated isotope-induced spectral shifting of hyperbolic phonon polaritons in hBN, the prior van der Waals example this three-dimensional crystal result extends.","marker":"43"}],"fun_headline_variants":["Isotope swap shifts shear polaritons 40 cm-1 lower","18O tunes gallium oxide polaritons into new band","Isotopic substitution opens previously inaccessible polariton range","Shear polaritons redshifted 40 cm-1 via oxygen isotope swap","Isotope substitution tunes hyperbolic shear polaritons"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing assumption is that the 1.2-micrometer-thick 18O epilayer behaves as a semi-infinite crystal for the launched polaritons, so the near-field images carry no contribution from the underlying 16O substrate.","fun_headline_variants_meta":{"raw":{"variants":["Isotope swap shifts shear polaritons 40 cm-1 lower","18O tunes gallium oxide polaritons into new band","Isotopic substitution opens previously inaccessible polariton range","Shear polaritons redshifted 40 cm-1 via oxygen isotope swap","Isotope substitution tunes hyperbolic shear polaritons"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000331,"raw_usage":{"total_tokens":1918,"prompt_tokens":1094,"completion_tokens":824,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":710,"completion_tokens_details":{"reasoning_tokens":735}},"tokens_in":710,"tokens_out":824,"duration_ms":8189,"temperature":1.0,"reasoning_tokens":735,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T13:09:44.431326+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Thin the 18O epilayer to about half a micrometer, or excite the same film with a larger antenna that launches lower momenta, and check whether the apparent 40 cm$^{-1}$ shift splits into two bands or moves toward the 16O substrate response; the semi-infinite assumption predicts no substrate contribution, while a substrate-coupling alternative predicts a mixed or thickness-dependent shift.","supporting_citations":[],"review_version":1}