{"id":"7341f327-eb20-4e30-bcb8-05e918ec5769","arxiv_id":"2411.10196","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":5.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"CVD-grown Er:Y2O3 films show substrate-dependent texture, including epitaxy on YSZ and 99% (111) texture on an MBE oxide/Si template, with telecom inhomogeneous linewidths of 9 to 14 GHz.","lead":"Researchers grew erbium-doped yttrium oxide films on six different substrates using chemical vapour deposition and found that the substrate controls film orientation, with epitaxial growth on yttria-stabilized zirconia and highly textured growth on an oxide-buffered silicon wafer. The erbium ions emit light in the telecom band with linewidths of 9 to 14 GHz, a useful step toward integrating quantum emitters with silicon photonics.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Unmeasured Er concentration confounds the central comparative claim that emission properties are substrate-independent.","rationale":"The reader's weakest_assumption identified exactly this issue: the Er3+ concentration is assumed to equal the nominal 670 ppm without direct measurement, and the authors note that doping differences could explain lifetime variations. I agree this is the most load-bearing concern. It is real but not fatal: the paper's claims are modest and appropriately hedged (e.g., 'comparable' rather than 'identical,' and explicit mention of doping as a possible explanation), the structural claims (epitaxy on YSZ, texture on MBE template) are supported by standard XRD evidence, and the optical data show qualitatively similar spectra and linewidths in the 9–19 GHz range. The lack of composition measurement does not invalidate the paper; it limits the strength of the comparative conclusion. Since the reader already accepted the paper with this caveat, my stress-test does not change the verdict. A concrete concentration measurement would settle whether the concern actually lands in the sense of altering the comparison's meaning.","tokens_in":9819,"tokens_out":6125,"duration_ms":60183,"concrete_test":"Measure the Er concentration in each as-grown film using Rutherford backscattering spectrometry (RBS) or secondary ion mass spectrometry (SIMS) on the films grown on Si(100), Si(111)/MBE-Y2O3, YSZ, sapphire, and quartz. If the measured concentrations agree with the nominal 670 ppm within analytical uncertainty (say ±20%) and are mutually consistent, the comparative optical claims are supported. If the concentrations vary by more than ~1.5× between samples, the claimed substrate-independence of emission properties is confounded and the comparison should be repeated on films with deliberately matched and independently verified concentrations.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's headline conclusion — that Er3+ emission properties are comparable across all substrates, indicating a high-quality local crystalline environment regardless of the substrate — is a comparative claim that implicitly assumes all films have the same Er3+ concentration. The nominal doping is stated as 670 ppm from the precursor (Table 1), but no composition measurement (RBS, SIMS, ICP-MS, XRF) is reported. If the actual incorporated Er concentration differs between films, then the observed differences in lifetimes (Fig. 6) and inhomogeneous linewidths (Table 1) could be dominated by doping-dependent effects (e.g., cross-relaxation, strain-induced disorder) rather than by the substrate-controlled film quality. Significantly, the authors themselves acknowledge this confound in Section 4: 'Small differences in the effective Er3+ doping level between samples could also partly explain the lifetime differences observed [18],' yet they do not quantify the doping or attempt to control for it. This is the single most load-bearing concern because the central claim is explicitly about comparability across substrates, and an uncontrolled variable directly affecting the measured observables prevents a clean interpretation of that comparison.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports the growth of Er:Y2O3 thin films by direct liquid injection CVD on a range of substrates: Si(100), Si(111), Si(111) with an MBE-grown Y2O3 template, quartz, sapphire, and YSZ(001). Structural characterization using SEM, AFM, and XRD shows substrate-dependent morphologies and textures, with fully (100)-oriented growth on YSZ, claimed to be epitaxial with a cube-on-cube relationship, and 99% (111) texture on the MBE template. Optical characterization at cryogenic temperatures shows visible and infrared Er3+ emission features that are qualitatively similar across all samples, with lifetimes of about 160 µs (4S3/2) and 8 ms (4I13/2), and inhomogeneous linewidths of 9-19 GHz for the telecom transition in the three samples examined at high resolution. The paper concludes that the emission properties are comparable across substrates, indicating a high-quality local crystal environment regardless of the substrate.","tokens_in":10157,"tokens_out":7613,"duration_ms":79550,"significance":"The paper demonstrates that a scalable and low-cost deposition technique (CVD) can produce rare-earth-doped oxide films with narrow inhomogeneous linewidths (9-14 GHz for the C2 site) on multiple substrates, including silicon via a hybrid MBE template. If the central comparability claim holds, this opens a practical route toward wafer-scale integration of Er-based quantum memories and telecom photon sources. The paper is clearly written, presents well-documented structural and optical data with reported fit uncertainties, and includes honest discussion of the remaining limitations. The main strength is the systematic side-by-side comparison of film quality across substrates, benchmarked against ceramic and MBE references.","major_comments":[{"comment":"The central claim that Er3+ emission properties are comparable across substrates implicitly assumes a constant Er3+ concentration in all films. The nominal concentration (670 ppm) is listed in Table 1, but no composition measurement (e.g., RBS, SIMS, XRF) is reported to verify the actual incorporated Er content. The authors themselves state in the Fig. 6 discussion that 'small differences in the effective Er3+ doping level between samples could also partly explain the lifetime differences observed [18].' Since lifetimes and inhomogeneous linewidths are known to depend on rare-earth concentration, the observed comparability could partly be a consequence of undetermined concentration variations rather than the local crystalline environment. I recommend measuring the Er content in each film or, if this is not feasible, explicitly reframing the conclusion as conditional on the as-assumed doping level and discussing the expected sensitivity of the observables to concentration.","section":"Section 4, Fig. 6, Table 1"},{"comment":"The 'epitaxial' descriptor used for the Y2O3 film on YSZ is supported by θ-2θ scans, a rocking curve with FWHM of 0.7°, and off-axis phi scans over a 0-180° azimuthal range. However, a full 360° phi scan and/or a cross-sectional TEM measurement would considerably strengthen this claim by excluding the presence of 90°-rotated domains or an amorphous interfacial layer. Without such data, the statement in the abstract and conclusions that the film is 'epitaxial' is somewhat stronger than the presented evidence strictly warrants.","section":"Section 3, Fig. 4"}],"minor_comments":[{"comment":"The lifetimes are reported as approximate values ('about 160 µs and 8 ms') without per-sample uncertainties or the number of repeated measurements. Since the comparability of lifetimes is part of the main conclusion, reporting uncertainties and measurement statistics would make the comparison more transparent.","section":"Fig. 6 and Section 4"},{"comment":"The texture metric defined by Eqs. (1) and (2) is empirical and not a standard quantitative texture coefficient. It would be helpful to state explicitly what physical quantity this percentage corresponds to (e.g., deviation from random orientation) and to note its limitations.","section":"Eqs. (1) and (2), Section 3"},{"comment":"The note that 'sample thickness was estimated from ellipsometric measurement performed on a film grown on a silicon substrate simultaneously' is ambiguous. It is unclear whether a Si witness piece was co-loaded in each deposition run or whether a single calibration was used for all substrates. Clarifying this would help assess the reliability of the reported thicknesses.","section":"Table 1"},{"comment":"The interpretation of the annealing results, namely 'diffusion of species between the film and the substrate' or 'stress release', is speculative and not directly supported by the presented data. Adding evidence such as SIMS depth profiles or XRD stress measurements, or presenting the interpretation as tentative, would strengthen this section.","section":"Section 4, Fig. 8"},{"comment":"There is a typo in the acknowledgements: 'Ackowledgements' should be 'Acknowledgements'. Also, references [2] and [7] are cited as arXiv preprints; if they have been published in peer-reviewed venues, updating the citations would be beneficial.","section":"General"}],"recommendation":"major_revision","confidential_remarks":"The concentration issue is the main technical concern, but it is addressable either by a straightforward RBS/SIMS measurement (which the authors' laboratories likely have access to) or by tempering the 'regardless of the substrate' claim. The paper is otherwise strong in its experimental documentation and honest about remaining limitations. The epitaxy evidence is solid but could be strengthened with a full phi scan or TEM. On balance, I believe the manuscript merits publication after these revisions."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Read this when you want to know where Er:Y2O3 thin films stand for quantum integration: CVD works, the lines are 9–14 GHz wide, and the hybrid MBE-template on Si gives 99% (111) texture. That's the new stuff. The comparison across substrates is the selling point, but it is a bit softer than the abstract implies.\n\nThe paper does what a good materials report should: standard XRD, SEM, AFM, PLE, with fit errors on the linewidths, and honest discussion of the texture metric and the annealing results. The authors compare to ceramics and MBE films rather than only to themselves. No derivations, no code, no circularity burden—the texture equations are just empirical ratios.\n\nThe main soft spot is real but not load-bearing: Er concentration is nominal only. The authors say in Section 4 that doping differences could explain lifetime variations, but they don't measure it, so the \"comparable emission properties\" claim is not as clean as presented. However, the linewidth numbers are direct measurements per sample; the range 9–14 GHz stands regardless of doping. Minor gaps: lifetimes without per-sample errors, phi scans over 0–180° only, no TEM for the epitaxial relationship on YSZ. These are small.\n\nBottom line: this is a solid, honest stepping stone. It doesn't claim more than it shows. For people working on rare-earth oxide films or CVD growth for quantum photonics, it's worth a careful read. Send it to peer review; a referee should push for composition measurements or a more guarded comparability statement, but the data is worth publishing.","headline":"Solid CVD materials report with a genuinely useful hybrid MBE-template route, undercut by unmeasured Er content but not fatally.","tokens_in":10593,"tokens_out":2208,"would_cite":true,"duration_ms":23577,"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":"Chemical vapour deposition can grow erbium-doped yttria films whose telecom emission is comparable across very different substrates, including an epitaxial film on YSZ and a highly textured film on silicon.","keywords":["erbium-doped yttrium oxide","chemical vapour deposition","thin films","quantum technologies","telecom wavelength","inhomogeneous linewidth","silicon integration","rare-earth ions"],"falsifier":"Measure the actual erbium concentration in each film, for example by Rutherford backscattering spectrometry or secondary-ion mass spectrometry, and correlate it with the measured lifetimes; if the concentrations differ substantially across substrates, the claim that the emission properties are substrate-independent would need to be revised.","tokens_in":9661,"feed_emoji":"📡","tokens_out":9358,"duration_ms":77788,"temperature":0.7,"pith_summary":"This paper tries to establish that chemical vapour deposition (CVD) — a scalable and flexible growth method — can produce erbium-doped yttrium oxide thin films good enough for quantum technologies, even on silicon substrates where oxide epitaxy is usually blocked by an amorphous interfacial layer. The authors grow films on silicon, sapphire, quartz, and yttria-stabilized zirconia, and on a silicon wafer carrying a molecular-beam-epitaxy yttria template. They find comparable erbium emission across all substrates, with an epitaxial film on YSZ and a 99% (111)-textured film on the silicon template. The telecom transition shows inhomogeneous linewidths of 9 to 14 GHz for the C2 site, close to MBE-grown films and far from the best ceramics. If the claim holds, CVD becomes a practical route for making the rare-earth-doped oxide layers that quantum memories and repeaters need, with a clear path to silicon integration.","feed_headline":"CVD grows epitaxial erbium-doped yttria on YSZ and silicon","feed_subtitle":"Telecom linewidths of 9 to 14 GHz persist across substrates, a step toward silicon-integrated quantum devices.","key_machinery":"The central object is the $\\mathrm{Er}^{3+}$ ion sitting on the two yttrium crystallographic sites (C2 and C3i) of cubic $\\mathrm{Y}_2\\mathrm{O}_3$, used as a local probe of crystalline quality. The $^4I_{13/2} \\rightarrow ^4I_{15/2}$ transition near 1536 nm is the telecom-relevant transition, and its inhomogeneous linewidth measured by photoluminescence excitation at 2.9 K quantifies static disorder around the ions. The growth machinery is direct liquid injection CVD, with a set of substrates chosen to test whether the amorphous $\\mathrm{SiO}_2$ interlayer on silicon can be circumvented by an MBE-grown $\\mathrm{Y}_2\\mathrm{O}_3$ template.","core_discovery":"The paper reports that direct liquid injection chemical vapour deposition can grow well-crystallized erbium-doped yttrium oxide ($\\mathrm{Er}{:}\\mathrm{Y}_2\\mathrm{O}_3$) films on silicon, sapphire, quartz, and yttria-stabilized zirconia (YSZ), and that the erbium emission is comparable on all of them despite very different film morphologies. On YSZ the film grows epitaxially with a cube-on-cube relationship, and on a molecular-beam-epitaxy-grown $\\mathrm{Y}_2\\mathrm{O}_3$ buffer on silicon it becomes 99% (111)-textured. Photoluminescence lifetimes of about 160 $\\mu$s (visible, 564 nm) and 8 ms (infrared, 1536 nm) are close to bulk ceramics, and the inhomogeneous linewidth of the telecom transition for erbium in the C2 site is 9–14 GHz. The authors interpret this as evidence that the local crystalline environment of $\\mathrm{Er}^{3+}$ is of high quality regardless of substrate, and they propose the CVD-on-template route as a path toward scalable silicon-integrated quantum devices.","pith_inferences":["A direct composition measurement of the erbium concentration in each film would test whether the substrate-independent lifetimes are real or an artifact of doping variation, which the paper leaves open.","If the emission properties really are independent of substrate, the same CVD approach could be tested with other rare-earth dopants, such as europium or ytterbium, for different quantum-relevant transitions.","The rough columnar morphology on YSZ may limit waveguide fabrication; growing at lower temperature or with different precursors might promote smoother layer-by-layer growth, but this is an extrapolation beyond the paper's data.","The measured linewidths are still tens of times broader than the best yttria ceramics, so the next step would be to identify whether strain or doping dominates the broadening by measuring films with varied erbium concentrations."],"forward_implications":["CVD can serve as a scalable deposition route for rare-earth-doped oxide films with telecom-relevant emission, without requiring MBE for the optically active layer.","The YSZ-grown film's cube-on-cube epitaxy makes it a candidate platform for coupling to optical cavities and microwave resonators, provided the surface roughness is addressed.","The 99% (111)-textured film on the MBE template shows that silicon integration is feasible by inserting an epitaxial oxide buffer, despite the native amorphous oxide issue.","The comparable linewidths across polycrystalline, textured, and epitaxial films imply that crystal orientation is not the limiting factor for inhomogeneous broadening in these films.","Annealing at 1000 °C broadens the lines, especially on silicon, so post-growth thermal processing should be avoided or re-optimized for this system."],"supporting_citations":[{"why":"Supplies the direct liquid injection CVD reactor and precursor method used for all films.","marker":"[12]"},{"why":"Provides the MBE-grown Er:Y2O3 on silicon benchmark for template growth and linewidth comparison.","marker":"[15]"},{"why":"Gives the MBE-grown Er:Y2O3 thin film lifetimes to which the CVD films are compared.","marker":"[7]"},{"why":"Reports the subkilohertz optical linewidths in Er:Y2O3 ceramics used as the best-case comparison.","marker":"[21]"},{"why":"Supplies nanoscale rare-earth optical coherence data used to contextualize the linewidth results.","marker":"[22]"},{"why":"Used to attribute lifetime differences to erbium doping level and non-radiative channels.","marker":"[18]"},{"why":"Documents interfacial reactions and diffusion in similar oxide films, relevant to the annealing results.","marker":"[13]"},{"why":"Explains the direct liquid injection CVD technique that underpins the growth method.","marker":"[14]"}],"fun_headline_variants":["CVD grows epitaxial Er:Y2O3 on YSZ and textured on Si","Telecom linewidths in CVD-grown Er:Y2O3 on multiple substrates","CVD-grown Er:Y2O3 films show 9-14 GHz linewidths on any substrate","Er:Y2O3 quantum films grown by CVD on silicon, YSZ, and more","CVD yields high-quality Er:Y2O3 films for quantum devices"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The paper assumes the erbium concentration is 670 ppm in every film, equal to the nominal precursor value, because no composition measurement is reported; if doping varies from film to film, the claimed substrate-independence of the optical properties could be an artifact of that variation.","fun_headline_variants_meta":{"raw":{"variants":["CVD grows epitaxial Er:Y2O3 on YSZ and textured on Si","Telecom linewidths in CVD-grown Er:Y2O3 on multiple substrates","CVD-grown Er:Y2O3 films show 9-14 GHz linewidths on any substrate","Er:Y2O3 quantum films grown by CVD on silicon, YSZ, and more","CVD yields high-quality Er:Y2O3 films for quantum devices"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000748,"raw_usage":{"total_tokens":3388,"prompt_tokens":1055,"completion_tokens":2333,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":671,"completion_tokens_details":{"reasoning_tokens":2216}},"tokens_in":671,"tokens_out":2333,"duration_ms":14947,"temperature":1.0,"reasoning_tokens":2216,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T19:50:59.434908+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the actual erbium concentration in each film, for example by Rutherford backscattering spectrometry or secondary-ion mass spectrometry, and correlate it with the measured lifetimes; if the concentrations differ substantially across substrates, the claim that the emission properties are substrate-independent would need to be revised.","supporting_citations":[{"cited_title":"Harada, A","cited_arxiv_id":null,"evidence_quote":"Supplies the direct liquid injection CVD reactor and precursor method used for all films."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the MBE-grown Er:Y2O3 on silicon benchmark for template growth and linewidth comparison."},{"cited_title":"Indistinguishable telecom band photons from a single erbium ion in the solid state","cited_arxiv_id":"2301.03564","evidence_quote":"Gives the MBE-grown Er:Y2O3 thin film lifetimes to which the CVD films are compared."},{"cited_title":"Singh, A","cited_arxiv_id":null,"evidence_quote":"Reports the subkilohertz optical linewidths in Er:Y2O3 ceramics used as the best-case comparison."},{"cited_title":"Bachelet, D","cited_arxiv_id":null,"evidence_quote":"Supplies nanoscale rare-earth optical coherence data used to contextualize the linewidth results."},{"cited_title":"Balașa, M.A","cited_arxiv_id":null,"evidence_quote":"Used to attribute lifetime differences to erbium doping level and non-radiative channels."},{"cited_title":"Dual epitaxial telecom spin-photon interfaces with correlated long-lived coherence","cited_arxiv_id":"2310.07120","evidence_quote":"Documents interfacial reactions and diffusion in similar oxide films, relevant to the annealing results."},{"cited_title":"Dibos, M","cited_arxiv_id":null,"evidence_quote":"Explains the direct liquid injection CVD technique that underpins the growth method."}],"review_version":1}