{"id":"2402f75e-4821-4e21-9fcc-59109266661d","arxiv_id":"2607.24685","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":5.5,"correctness_risk":"low","formal_verification":"none","parameter_count":4,"one_line_summary":"Epitaxial planar hexagonal Ge forms on m-plane CdS by LEPECVD at ~250 °C, with Raman E2g selection rules and rapid near-interface strain relief, but hexagonal order collapses beyond ~6–12 nm via I3 stacking faults and no Ge photoluminescence is seen.","lead":"Researchers grew thin planar films of hexagonal germanium on cadmium sulfide using a low-temperature plasma CVD method and confirmed the crystal phase with microscopy, X-rays, and polarized Raman light. Planar hexagonal Ge is sought for silicon-compatible photonics and spin devices, but the pure hexagonal order only lasts a few nanometers before stacking faults take over.","discovery_kind":"extension","skeptic_critique":{"model":"moonshotai/kimi-k3","headline":"The XRD-derived ~2:1 2H:3C volume ratio likely conflates I3/stacking-disordered material with well-ordered Ge-2H, since STEM shows extended 2H domains nearly vanish beyond ~12 nm while cubic domains never dominate; the \"2H fraction\" is therefore a lower-quality-of-evidence quantity than the paper'sR","rationale":"The reader identified the Raman E2g assignment/subtraction as the weakest assumption; I find that concern secondary because the 287–290 cm⁻¹ band is spectrally well separated from the known confounders (Ge-3C at 300.3, a-Ge at 275, CdS modes) and, crucially, the angular two-fold modulation of that band (Fig. 4(d)) is an independent symmetry test that a substrate/cubic superposition would not reproduce at that frequency. The more load-bearing quantitative issue is the XRD phase-fraction estimate in §III.A, which the reader cited but treated as consistent context (\"~2:1 2H:3C volume ratio\") rather than as itself soft: it is derived from a single peak pair while the authors' own STEM analysis shows most of the 50 nm film is stacking-disordered material that would diffract into the 2H channels. This inflates the apparent well-ordered hexagonal fraction and is the quantity most likely to mislead a reader planning follow-up optical or transport work. I nonetheless recommend UNCHANGED (ACCEPT) rather than CONDITIONAL, because (a) the existence claim of epitaxial Ge-2H at the interface rests on three independent techniques and is not threatened, (b) the authors themselves disclose the 6–12 nm loss of hexagonal order and the absence of Ge PL, so the overstatement is one of emphasis rather than of evidence, and (c) the proposed test is a refinement of quantification, not a check that could overturn the central claim.","tokens_in":17395,"tokens_out":3208,"duration_ms":127144,"concrete_test":"Two-part check on sample IV. (1) STEM census: count stacking sequences (clean 2H ≥2 AB pairs vs I3 vs 3C vs unresolvable) across at least three FIB lamellae from laterally separated regions, integrate the clean-2H fraction over the full 50 nm thickness, and compare to the XRD 2:1 estimate; if the integrated well-ordered 2H fraction is ≪2/3, the phase-purity claim needs rephrasing. (2) XRD line-profile analysis: fit the (2̄201) and (2̄310) peak widths along q∥ and q⊥ against instrument resolution, and record a fault-sensitive reflection (e.g., a (10̄1l)-type scan along c*) to look for stacking-fault streaking; a Warren-type fault-density fit would separate coherent 2H volume from faulted volume and test whether the \"2H\" intensity is inflated.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The phase-existence claim (planar Ge-2H stabilized at 250 °C) is well supported: STEM ABAB stacking at the interface, distinct Ge-2H (2̄201)/(2̄310) RSM peaks, and the two-fold E2g polarization modulation at 287–290 cm⁻¹ (a frequency cleanly separated from Ge-3C F2g at 300.3 cm⁻¹, a-Ge at 275 cm⁻¹, and CdS modes) all point the same way. The soft spot is quantitative, and it sits in §III.A rather than in the Raman section. From the integrated intensities of exactly one peak pair — Ge-2H (2̄201) vs Ge-3C (1̄1̄3), with xrayutilities structure factors — the authors estimate \"a roughly 2:1 ratio of 2H and 3C phases\" in the 50 nm film. But their own STEM census of the same sample (Fig. 2(a)) says that beyond ≈12 nm, extended hexagonal domains are \"nearly absent,\" cubic domains \"do not dominate either,\" and the remaining volume is \"largely governed by I3 defects or more general stacking disorder.\" I3-faulted and randomly faulted material still contributes diffracted intensity at or near the 2H Bragg positions (stacking faults along c mainly produce streaking/broadening along fault-sensitive reciprocal directions, not intensity at the 3C positions). So the XRD \"2H\" integral almost certainly counts faulted, non-2H-ordered volume as 2H, while the \"3C\" integral counts only well-ordered cubic inclusions. The true well-ordered 2H volume fraction in sample IV may be far below the implied ~2/3 — consistent with STEM, where clean 2H is confined to the first ~5–6 nm of 50 nm. This does not break the central stabilization claim (the near-interface phase is unambiguous), but it weakens the implicit message that the film is predominantly hexagonal-ordered, and it affects any downstream use of this material (e.g., interpreting the absent PL, or planning optical/transport measurements that assume a bulk-like 2H layer). The reader's flagged Raman-subtraction issue is real but less load-bearing, because the E2g band is spectrally isolated and its selection-rule modulation is independent of","agreement_with_reader":"partial"},"referee_report":{"model":"moonshotai/kimi-k3","summary":"The manuscript reports LEPECVD growth of Ge on commercial non-basal m-plane CdS and identifies 250 °C as an intermediate growth window between amorphous deposition at 200 °C and interfacial degradation at 300 °C. HAADF-STEM along two zone axes shows interfacial ABAB stacking, cubic inclusions, and I3-type basal faults; asymmetric RSMs resolve Ge-2H reflections and provide lattice parameters; polarization-resolved Raman spectra show the expected E2g selection rules; and EELS limits appreciable intermixing to about 2 nm. No Ge-related PL is detected. The authors further identify two dislocation families consistent with rapid anisotropic strain relaxation and describe a transition from relatively well-ordered interfacial Ge-2H to stacking-disordered material beyond roughly 6–12 nm.","tokens_in":27226,"tokens_out":1059,"duration_ms":477863,"significance":"If the interpretation holds, this is a useful advance toward planar, potentially scalable hexagonal-Ge heterostructures rather than core–shell nanowires. The phase assignment is falsifiable and is supported by several largely independent channels: atomic-resolution stacking along two projections, two asymmetric XRD geometries, polarization selection rules for a spectrally separated E2g mode, and chemical intermixing measurements. The temperature series and defect analysis are also valuable for understanding why hexagonal order is lost with thickness. The concurrent MBE work of Koolen et al. is appropriately acknowledged, so the novelty is primarily the LEPECVD route and the detailed relaxation/defect characterization rather than the first planar Ge/CdS report. The public data availability statement is a further strength.","major_comments":[{"comment":"The estimated “roughly 2:1 ratio of 2H and 3C phases” should not be presented as a well-ordered volume fraction without further qualification. It is obtained from one integrated peak pair, while the STEM census of the same sample states that extended 2H domains are nearly absent beyond ~12 nm and that much of the remaining volume is I3-faulted or otherwise stacking disordered. Such faulted material can still contribute intensity near the 2H Bragg position, whereas the 3C peak preferentially counts ordered cubic material. The integration ranges, peak widths, grazing-incidence footprint/absorption corrections, texture effects, and uncertainty are also not reported. Please recast this as an apparent Bragg-intensity ratio, provide the needed corrections and uncertainty, or reconcile it quantitatively with the STEM census. This does not undermine the interfacial Ge-2H identification, but it d","section":"§III.A, Fig. 3(a)"}],"minor_comments":[{"comment":"The conversion from the measured (2-310) peak to a_parallel is unclear. The text says a_parallel is obtained through the (-1-120) spacing d_a = a_parallel/2, but the reported values appear to use a_parallel ≈ 1/q_parallel (1/2.5148 = 0.398 nm). Please define the reciprocal-space components and the conversion explicitly.","section":"§III.A, discussion of Fig. 3(b)"},{"comment":"The statement that the modeled dislocations are “energetically viable” is stronger than the reported tens-of-picoseconds MD stability test. The simulations demonstrate kinetic persistence under the chosen potential and boundary conditions, not relative thermodynamic stability. Please either qualify the wording or provide defect formation/relative energies and a MatterSim benchmark for relevant Ge-2H stacking-fault or dislocation configurations.","section":"§III.C"},{"comment":"Please state how the spectra in Fig. 4(a) were normalized and how the confidence band in Fig. 4(d) was calculated. It would also help to specify whether the stated ~7.5° azimuthal offset was included as a fitted parameter or fixed after alignment.","section":"§III.B, Fig. 4"},{"comment":"Because the text draws a conclusion from the significantly different PL intensities of sample IV and bare CdS, the caption should state whether the plotted intensities are absolute or rescaled and whether collection geometry, substrate thickness, and acquisition parameters were identical.","section":"§III.B, Fig. 5"},{"comment":"Minor typographical errors in the characterization section: “perfomed” should be “performed,” and “operation at 20 kV” should be “operating at 20 kV.”","section":"§II.B"}],"recommendation":"minor_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"The punchline is straightforward: they stabilize epitaxial planar hexagonal Ge on m-plane CdS by LEPECVD in a narrow window around 250 °C, and they map how it relaxes and then disorders with thickness. Concurrent MBE work on the same template is cited, so novelty is shared, but the LEPECVD temperature window, polarization-resolved E2g selection rules, residual lattice parameters from two RSMs, and the combined STEM + atomistic dislocation/I3 analysis are real independent value.\n\nWhat they do well is the multi-probe consistency. HAADF-STEM along [11-20] and [0001] shows clear ABAB stacking near the interface; RSMs give distinct Ge-2H (2-201) and (2-310) peaks with lattice constants close to the nanowire literature; the E2g band at 287–290 cm⁻¹ is spectrally clean of Ge-3C F2g and a-Ge and follows the expected two-fold polarization; EELS limits intermixing to ~2 nm; 300 °C produces the same mushroom/intermix degradation seen in the MBE paper. The dislocation models (partial ½[0001]-type with cubic insertions for c-relaxation, a-type ⅓[11-20] for a-relaxation) plus MatterSim MD stability checks match the TEM contrast and explain why most mismatch is gone within a few nm. Data are open. That is careful experimental materials work.\n\nThe soft spot is quantitative, not existential. STEM on the 50 nm film shows extended 2H domains nearly gone past ~12 nm, cubic never dominant, and the rest I3 or general stacking disorder. XRD’s “roughly 2:1 2H:3C” from one peak pair almost certainly folds faulted volume into the 2H integral (faults streak/broaden rather than move intensity to 3C positions). So the well-ordered 2H fraction is thinner than the ratio implies—consistent with their own images and with the missing Ge PL. Raman subtraction of the mixed ~302 cm⁻¹ band is a lesser issue; the E2g fingerprint itself holds. Authors already state the thickness and PL limits; they just do not fully reconcile the XRD fraction with the STEM census.\n\nThis is for people working on hexagonal group-IV epitaxy, templates, and defect engineering. It will not give you thick phase-pure material or optical emission yet, but the structural claim and the defect mechanisms are solid enough to build on. I would send it to referees; minor revision on how they phrase the volume fractions is enough. Engage with it if you care about planar hex-Ge routes.","headline":"Solid multi-technique demonstration of planar Ge-2H on m-CdS by LEPECVD, with a useful defect map; the main soft spot is that the XRD ~2:1 2H:3C ratio overstates well-ordered hexagonal volume once I3 disorder sets in.","tokens_in":18551,"tokens_out":661,"would_cite":true,"duration_ms":12698,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.5","headline":"Planar hexagonal germanium grows epitaxially on m-plane CdS near 250 °C, with most mismatch strain relieved in a few nanometers before stacking faults erode long-range 2H order.","keywords":["hexagonal germanium","Ge-2H","CdS template","m-plane epitaxy","LEPECVD","stacking faults","strain relaxation","polarized Raman"],"falsifier":"A thickness- or process-series experiment that yields a continuous >20 nm film whose XRD hexagonal-to-cubic intensity ratio approaches pure 2H, whose STEM [11̄20] maps show extended ABAB domains free of I3 faults, and whose co-polarized Raman E2g intensity follows the full angular selection rule without residual intensity assignable to cubic Ge or CdS.","tokens_in":18144,"feed_emoji":"💎","tokens_out":1081,"duration_ms":22673,"temperature":0.7,"pith_summary":"Metastable hexagonal (lonsdaleite) germanium is prized for a possibly direct or quasi-direct gap and a large Landé g-factor, but until now it has mostly been locked into nanowire shells that are hard to integrate. This work shows that non-basal m-plane CdS can template continuous planar Ge-2H by low-energy plasma-enhanced CVD when the growth temperature sits near 250 °C: hot enough for adatom mobility, cool enough to avoid a displacement reaction that destroys the interface. STEM shows the expected ABAB stacking next to the interface, XRD finds Ge-2H Bragg peaks with residual strain close to nanowire values, and the E2g Raman mode follows hexagonal polarization selection rules. Strain is dumped quickly by a short list of dislocations—some of which insert thin cubic slabs—while farther out I3 basal stacking faults scramble the hexagonal registry. No clear Ge photoluminescence appears, consistent with a weak pseudo-direct gap and the limited pure-hexagonal thickness. The result positions CdS as a workable planar template and maps the defect ladder that must be tamed if thicker, phase-pure films are to be useful.","feed_headline":"Planar hexagonal Ge grows on CdS before stacking faults take over","feed_subtitle":"At 250 °C, m-plane CdS locks in Ge-2H for a few nanometers; dislocations then cubic slabs and I3 faults erode order.","key_machinery":"m-plane CdS templating of ABAB stacking: the non-basal (1̄100) surface preserves the hexagonal bilayer sequence and kinetically frustrates cubic ABCABC stacking, enabling planar Ge-2H when growth temperature balances adatom mobility against thermochemical Ge–CdS intermixing.","core_discovery":"Epitaxial planar Ge-2H can be stabilized on m-plane CdS by LEPECVD at about 250 °C. Within the first few nanometers the film is predominantly hexagonal; a limited set of misfit dislocations relieves most of the anisotropic ~2–3.6 % mismatch (sometimes inserting localized cubic stacking), after which I3-type stacking-fault disorder progressively destroys long-range 2H order. Polarization-resolved Raman confirms the E2g selection rules of the hexagonal phase, while photoluminescence shows no Ge-related emission.","pith_inferences":["If I3 faults are intrinsic to Ge-2H rather than mismatch-driven, alloying (e.g., SiGe-2H) or surfactant-mediated growth may be required even on lattice-matched templates.","The same limited dislocation set and cubic-slab insertion pathway may appear in other metastable hexagonal group-IV / II–VI stacks, offering a transferable defect taxonomy.","Absence of PL here does not rule out a quasi-direct gap; resonant or cavity-enhanced measurements on sub-10 nm pure-hexagonal slabs could still reveal weak emission once substrate background is removed."],"forward_implications":["CdS m-plane wafers become a practical planar platform for hexagonal group-IV epitaxy instead of only III–V nanowire cores.","Device-relevant Ge-2H layers will be limited to roughly the first 6–12 nm unless growth is retuned to suppress I3 basal stacking faults after plastic relaxation.","Strain engineering and dislocation engineering (a-type and dissociated c-component partials) are the primary knobs for residual lattice constants and phase purity.","Optical characterization of bulk-like planar Ge-2H still requires thicker, cleaner films before emission can be cleanly separated from CdS sub-gap bands."],"fun_headline_variants":["CdS locks planar Ge-2H for a few nm before I3 faults erode order","Epitaxial hexagonal Ge on m-plane CdS stabilizes then yields to stacking faults","LEPECVD grows Ge-2H on CdS; dislocations and I3 defects end long-range order","Planar Ge-2H on CdS: Raman confirms E2g, order fades past misfit dislocations","m-plane CdS templates hexagonal Ge until cubic slabs and I3 faults dominate"],"cache_read_input_tokens":128,"weakest_assumption_plain":"That the Raman band near 287–290 cm⁻¹ is a clean E2g fingerprint of hexagonal Ge, even though the neighboring ~302 cm⁻¹ band mixes hexagonal, cubic, and substrate signals and XRD already finds a substantial cubic volume fraction in thicker films.","fun_headline_variants_meta":{"raw":{"variants":["CdS locks planar Ge-2H for a few nm before I3 faults erode order","Epitaxial hexagonal Ge on m-plane CdS stabilizes then yields to stacking faults","LEPECVD grows Ge-2H on CdS; dislocations and I3 defects end long-range order","Planar Ge-2H on CdS: Raman confirms E2g, order fades past misfit dislocations","m-plane CdS templates hexagonal Ge until cubic slabs and I3 faults dominate"]},"model":"grok-4.5","effort":"low","cost_usd":0.003835,"raw_usage":{"total_tokens":1283,"prompt_tokens":857,"num_sources_used":0,"completion_tokens":106,"cost_in_usd_ticks":38348000,"prompt_tokens_details":{"text_tokens":857,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":320,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":857,"tokens_out":106,"duration_ms":7323,"temperature":1.0,"reasoning_tokens":320,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-31T07:51:34.066608+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"A thickness- or process-series experiment that yields a continuous >20 nm film whose XRD hexagonal-to-cubic intensity ratio approaches pure 2H, whose STEM [11̄20] maps show extended ABAB domains free of I3 faults, and whose co-polarized Raman E2g intensity follows the full angular selection rule without residual intensity assignable to cubic Ge or CdS.","supporting_citations":[],"review_version":1}