{"id":"1aa0bc93-8ca5-4361-b9d5-11212764d3d8","arxiv_id":"2607.10908","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":7,"one_line_summary":"Ge-doped (-201) β-Ga₂O₃ on 6° offcut c-sapphire by LPCVD delivers record 117 cm²/V·s RT mobility (337 at 128 K) among Ge-doped films on sapphire, with transport fits giving shallow donors ~12–19 meV and Ndis ~10⁹ cm⁻².","lead":"LPCVD-grown Ge-doped β-Ga₂O₃ films on 6° offcut sapphire reach 117 cm²/V·s room-temperature Hall mobility (337 cm²/V·s at 128 K), the highest reported for Ge-doped films on sapphire. This shows offcut engineering can push heteroepitaxy toward homoepitaxy-level transport on cheap, scalable wafers for power devices.","discovery_kind":"extension","skeptic_critique":{"model":"grok-4.5","headline":"Mobility trend is confounded by systematic co-variation of Ga–substrate distance, Ge loading, thickness and growth rate with offcut (Table 1), so the causal attribution to step-flow alone is not isolated.","rationale":"The measured Hall numbers and the “highest-reported-for-Ge-on-sapphire” statement are directly supported by the data and the literature comparison in Fig. 6; no fabrication or internal-inconsistency red flag appears. The single load-bearing weakness is exactly the one the reader flagged: co-variation of source geometry and doping with offcut angle prevents clean isolation of the step-flow mechanism. A fixed-geometry control series would settle the issue; until then the CONDITIONAL verdict (high confidence for an experimental materials paper) is appropriate and needs no adjustment.","tokens_in":21077,"tokens_out":619,"duration_ms":19883,"concrete_test":"Re-grow the exact low-doping recipe of Sample 4 (Ge 0.05 wt%, Ga–substrate distance fixed at 4.5 cm, target thickness ~3 µm) on 0° and 2° offcut sapphire under otherwise identical LPCVD conditions; if the room-temperature mobility remains near 117 cm² V⁻¹ s⁻¹ (or the rocking-curve FWHM and terrace order do not degrade), the offcut is not load-bearing; if mobility collapses toward the original 15–57 cm² V⁻¹ s⁻¹ values, the confound is confirmed.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper’s central causal claim—that sapphire offcut is the primary driver of the mobility jump from 15 to 117 cm² V⁻¹ s⁻¹—rests on the assumption that all other growth variables are held fixed or are secondary. Table 1 shows they are not: Ga–substrate distance is stepped from 2.5 cm (0°) to 5.5 cm (8°), Ge/(Ge+Ga) loading drops from 0.50–0.60 wt% (high-n, low-µ samples) to 0.05 wt% (the record 117 cm² V⁻¹ s⁻¹ sample), film thickness ranges 0.85–5.34 µm and growth rate more than triples. These parameters independently control Ge incorporation, residual impurity incorporation, strain relaxation and dislocation filtering. The structural data (SEM/AFM terrace alignment, rocking-curve FWHM) do improve with offcut, but the electrical optimum occurs precisely at the lowest doping and a unique source geometry; without a fixed-geometry control series the offcut attribution remains correlative rather than causal. The multi-parameter Boltzmann fit (different ħω₀ for 6° vs 8°, free ND1/ND2/NA/Ndis) further softens the mechanistic interpretation but is secondary to the experimental confound.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.5","summary":"The manuscript reports LPCVD heteroepitaxy of Ge-doped (−201) β-Ga₂O₃ on c-plane sapphire with intentional offcuts of 0°, 2°, 6°, and 8°. Structural data (SEM, AFM, XRD rocking curves, Raman) show a transition toward step-flow growth, improved terrace alignment, lower RMS roughness, and narrower (402) FWHM with increasing offcut. XPS indicates near-stoichiometric O/Ga ≈ 1.48. Hall measurements give room-temperature mobilities from 15 to 117 cm²/V·s over n ≈ 1.43×10¹⁷–2.75×10¹⁸ cm⁻³, with the 6° sample reaching 117 cm²/V·s at 1.43×10¹⁷ cm⁻³ and a peak of 337 cm²/V·s at 128 K. Temperature-dependent n(T) and μ(T) are fit with a two-donor charge-neutrality model plus Matthiessen scattering (POP, II, NI, ADP, dislocation), yielding ED ≈ 12.5–19 and 80 meV, low NA, and Ndis ~ 10⁹ cm⁻². The authors conclude that sapphire offcut engineering enables the highest reported RT and LT mobilities for Ge-doped β-Ga₂O₃ on sapphire and is a scalable path for high-mobility heteroepitaxy.","tokens_in":21474,"tokens_out":1668,"duration_ms":22515,"significance":"If the mobility values and structural trends hold, the work is a useful advance for scalable β-Ga₂O₃ heteroepitaxy: LPCVD Ge doping on low-cost sapphire reaching ~117 cm²/V·s approaches homoepitaxial Ge-doped ranges and substantially exceeds prior Ge-doped β-Ga₂O₃ on sapphire. The multi-technique characterization (phase purity, morphology, composition, T-dependent Hall) and standard transport modeling are appropriate for the field. The result is of practical interest for power-device drift layers on foreign substrates. The main scientific contribution is experimental demonstration and benchmarking rather than a new theoretical framework; its impact depends on how cleanly mobility gains can be attributed to offcut versus co-varying growth parameters.","major_comments":[{"comment":"Table 1 and §II–III: The central causal claim that sapphire offcut is the primary driver of the mobility rise (15 → 117 cm²/V·s) is not isolated. Ga–substrate distance is stepped 2.5→5.5 cm with offcut, Ge/(Ge+Ga) loading drops from 0.50–0.60 wt% (0°/2°) to 0.05 wt% for the record Sample 4, and thickness/growth rate vary by factors of ~3–5. These parameters independently control Ge flux, residual impurities, strain relaxation, and dislocation filtering. The electrical optimum coincides with the lowest intentional doping and a unique source geometry. Without fixed-geometry / fixed-loading controls (or at least a systematic discussion quantifying each contribution), the attribution to step-flow alone remains correlative. Please either (i) add control growths at fixed Ga–substrate distance and comparable n across offcuts, or (ii) substantially reframe Abstract/Conclusion language to “optimi","section":null},{"comment":"§III, Table 1, Fig. 7–8: Mobility comparisons across offcut are confounded by large differences in carrier concentration. The 0° and 2° films sit at n ~ (2.5–2.8)×10¹⁸ cm⁻³ while the 117 cm²/V·s film is at 1.43×10¹⁷ cm⁻³; Fig. 8 already shows that within 6° films, lowering n from 4.24×10¹⁷ to 1.43×10¹⁷ raises RT μ from 78 to 117 cm²/V·s. Ionized-impurity scattering therefore accounts for a substantial fraction of the reported “offcut” gain. Present mobility vs n (as in Fig. 6) with offcut as a parameter, and avoid ranking offcuts by raw μ without n-matching. Clarify which samples establish an offcut benefit at fixed doping.","section":null},{"comment":"§III, Eqs. (1)–(3), Tables 2–3, Fig. 9: The transport fit uses several free parameters (ND1, ND2, ED1, ED2, NA, Ndis, and different fitted ħω₀ of 47 vs 39 meV for 6° vs 8°). Different phonon energies for the same material under similar growth conditions need physical justification or a common ħω₀ with uncertainty bounds. Report fit sensitivity / uniqueness (e.g., covariance or constrained fits) so that extracted Ndis ~ 1–1.7×10⁹ cm⁻² and NA < 5×10¹⁵ cm⁻³ are not under-determined. The model is standard and secondary to the experimental claim, but over-interpreted uniqueness would weaken the mechanistic narrative.","section":null}],"minor_comments":[{"comment":"Notation: The orientation is written inconsistently as (2\"01), (−201), and (-201). Standardize to (−201) / (¯201) throughout text, abstract, figures, and tables.","section":null},{"comment":"Fig. 3(b): Rocking-curve FWHM values are shown only as a trend; please list numerical FWHM (arcsec) for each offcut in the text or a table so the crystallinity improvement can be cited quantitatively.","section":null},{"comment":"Table 1: Sample 7 lists thickness 5.34 µm and growth rate 3.56 µm/h (implying ~1.5 h), while several others equate thickness to growth rate numerically (suggesting 1 h growth). State growth time explicitly for each run.","section":null},{"comment":"Fig. 6 caption and legend: Ensure all literature points are correctly attributed (homoepitaxy vs heteroepitaxy; β vs α) and that the “highest on sapphire” claim is restricted to Ge-doped β-Ga₂O₃ as stated, not all Ga₂O₃:Ge.","section":null},{"comment":"XPS: O/Ga = 1.48 is near-stoichiometric; briefly note surface sensitivity and whether adventitious C / O–H affect the ratio, and whether Ge is detectable at the doping levels used (or below XPS detection).","section":null},{"comment":"§I: Prior LPCVD Si-doped work from the group is appropriately cited; a short explicit contrast (Ge vs Si incorporation efficiency under the same reactor) would help readers place the Ge results.","section":null},{"comment":"Typographical: scattered OCR-like artifacts in the provided text (e.g., 10!\" cm#$, µ;<;) should be cleaned in the production manuscript; ensure Greek and subscripts render correctly.","section":null}],"recommendation":"major_revision","confidential_remarks":"The absolute mobility numbers and structural trends look credible and of interest for APL Materials / similar venues. The main risk is overselling offcut as a single-variable lever when Table 1 shows a multi-parameter optimization. If the authors reframe honestly and tighten n-matched comparisons, this can become a solid contribution; if they insist on a pure offcut narrative without controls, the paper remains correlative. No integrity concerns; novelty relative to prior LPCVD Ge:Ga₂O₃ on sapphire (Ranga et al.) is real on the mobility numbers."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"The new experimental fact is solid: LPCVD Ge-doped (-201) β-Ga₂O₃ on 6° offcut sapphire reaches 117 cm²/V·s at 1.43×10¹⁷ cm⁻³ (RT) and 337 cm²/V·s at 128 K. That is well above prior Ge-on-sapphire LPCVD (~20–43) and approaches good homoepitaxial Ge values. Structural data (SEM/AFM terrace alignment, rocking-curve narrowing, Raman, near-stoichiometric XPS) are consistent and support better crystallinity at higher offcut.\n\nWhat the paper does well is the multi-offcut matrix plus temperature-dependent Hall and a standard charge-neutrality + Matthiessen fit (POP, II, NI, ADP, DIS). Extracted shallow donors (12.5–19 meV), a deeper ~80 meV level, low NA, and Ndis ~10⁹ cm⁻² are plausible and match literature ranges. The benchmark plot is honest.\n\nThe soft spot is real but not fatal. Table 1 shows Ga–substrate distance, Ge loading, thickness, and growth rate all step with offcut. The record sample is also the lightest-doped and uses a unique source geometry. Structural improvement with offcut is clear; the electrical jump is therefore correlative rather than cleanly isolated. The model also floats ħω₀ differently for 6° vs 8°. That weakens the “offcut is the primary knob” claim but does not erase the mobility numbers themselves.\n\nThis is for UWBG materials and device groups who need high-mobility n-type layers on cheap sapphire. Cite the Hall values and the growth recipe; treat the pure-offcut mechanism with the usual caution until a fixed-geometry control series appears. I would send it to peer review—the data are worth referee time and the confound is fixable by clarification or extra experiments.","headline":"Record Ge:β-Ga₂O₃/sapphire Hall numbers are real and useful; the offcut-only causal story is confounded by co-varying source geometry and doping.","tokens_in":22173,"tokens_out":500,"would_cite":true,"duration_ms":4707,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.5","headline":"Sapphire offcut of 6° yields Ge-doped β-Ga₂O₃ films with 117 cm²/V·s room-temperature mobility on low-cost substrates, the highest reported for this heteroepitaxy.","keywords":["β-Ga₂O₃","Ge doping","LPCVD","sapphire offcut","heteroepitaxy","Hall mobility","step-flow growth","ultra-wide bandgap"],"falsifier":"Grow a new set of films that hold Ga–substrate distance, Ge loading, thickness and growth rate fixed while varying only the sapphire offcut; if the mobility peak at 6° disappears, the offcut claim is falsified.","tokens_in":21995,"feed_emoji":"⚡","tokens_out":852,"duration_ms":6452,"temperature":0.7,"pith_summary":"The paper shows that intentional offcut on c-plane sapphire steers low-pressure chemical vapor deposition of Ge-doped β-Ga₂O₃ into a step-flow growth mode. That change improves terrace alignment, crystalline coherence and surface smoothness, which in turn raises electron mobility. On a 6° offcut the films reach 117 cm²/V·s at room temperature and 337 cm²/V·s at 128 K—numbers that surpass all earlier Ge-doped β-Ga₂O₃ layers grown on sapphire and approach values previously limited to expensive native substrates. Transport modeling extracts shallow donors (12.5–19 meV), a deeper 80 meV level, very low compensation and dislocation densities near 10⁹ cm⁻². The result matters because high-mobility n-type films on scalable sapphire open a practical route to ultra-wide-bandgap power devices without relying on costly bulk Ga₂O₃ wafers.","feed_headline":"6° sapphire offcut lifts Ge-Ga₂O₃ mobility to 117 cm²/V·s","feed_subtitle":"Highest room- and low-temperature values yet for Ge-doped β-Ga₂O₃ on low-cost sapphire","key_machinery":"Sapphire offcut engineering: intentional substrate miscut toward ⟨11-20⟩ that supplies atomic steps, stabilizes step-flow growth, suppresses rotational domains and thereby lowers the scattering that had previously limited mobility in heteroepitaxial Ge-doped β-Ga₂O₃.","core_discovery":"LPCVD growth of Ge-doped (-201) β-Ga₂O₃ on 6° offcut c-plane sapphire produces the highest room-temperature (117 cm²/V·s) and low-temperature (337 cm²/V·s) Hall mobilities yet reported for Ge-doped β-Ga₂O₃ on sapphire, by converting multidirectional island growth into ordered step-flow growth that reduces rotational domains and dislocation scattering.","pith_inferences":[],"forward_implications":[],"fun_headline_variants":["6° offcut sapphire yields Ge-β-Ga₂O₃ mobility of 117 cm²/V·s","Step-flow growth on 6° sapphire lifts Ge-Ga₂O₃ to 117 cm²/V·s","Ge-doped β-Ga₂O₃ on 6° sapphire hits 117 cm²/V·s Hall mobility","Highest Ge-Ga₂O₃ sapphire mobility: 117 cm²/V·s via 6° offcut","LPCVD on 6° offcut sapphire gives Ge-β-Ga₂O₃ 117 cm²/V·s"],"cache_read_input_tokens":16512,"weakest_assumption_plain":"The mobility gain is attributed mainly to the sapphire offcut angle, even though Ga-to-substrate distance, Ge source loading, film thickness and growth rate all change together with offcut in the experimental series.","fun_headline_variants_meta":{"raw":{"variants":["6° offcut sapphire yields Ge-β-Ga₂O₃ mobility of 117 cm²/V·s","Step-flow growth on 6° sapphire lifts Ge-Ga₂O₃ to 117 cm²/V·s","Ge-doped β-Ga₂O₃ on 6° sapphire hits 117 cm²/V·s Hall mobility","Highest Ge-Ga₂O₃ sapphire mobility: 117 cm²/V·s via 6° offcut","LPCVD on 6° offcut sapphire gives Ge-β-Ga₂O₃ 117 cm²/V·s"]},"model":"grok-4.5","effort":"low","cost_usd":0.007582,"raw_usage":{"total_tokens":1981,"prompt_tokens":1028,"num_sources_used":0,"completion_tokens":152,"cost_in_usd_ticks":75820000,"prompt_tokens_details":{"text_tokens":1028,"audio_tokens":0,"image_tokens":0,"cached_tokens":128},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":801,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":1028,"tokens_out":152,"duration_ms":6305,"temperature":1.0,"reasoning_tokens":801,"cache_read_input_tokens":128,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-14T08:21:55.881239+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"Grow a new set of films that hold Ga–substrate distance, Ge loading, thickness and growth rate fixed while varying only the sapphire offcut; if the mobility peak at 6° disappears, the offcut claim is falsified.","supporting_citations":[],"review_version":1}