{"id":"2bfdbe7d-2608-4ec0-b306-94496e55a544","arxiv_id":"2607.27424","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":5.0,"correctness_risk":"low","formal_verification":"none","parameter_count":3,"one_line_summary":"Stoichiometric low-index β-Ga2O3 surfaces dominate under normal growth conditions; Ga-rich adlayer terminations appear on (100) and (¯201) only under highly reducing conditions.","lead":"A DFT survey of all low-index β-Ga2O3 surfaces finds stoichiometric terminations stable across nearly all growth-relevant oxygen chemical potentials, with Ga-rich adlayer-like terminations only under extreme reducing conditions. The results give growers a practical map of which facets and terminations to expect when tuning oxygen pressure.","discovery_kind":"extension","skeptic_critique":{"model":"grok-4.5","headline":"No significant objection identified","rationale":"The paper delivers a clean, reproducible DFT surface-thermodynamics survey. The strongest claim is a direct readout of the computed γ(Δμ_O) diagrams under the bulk-truncated model the authors define; methods (all-electron FHI-aims, PBEsol vs PBE0(0.26), harmonic phonopy corrections, Wulff shapes) are standard and the data are deposited. The only material residual risk is precisely the one the reader names—missing global reconstruction/adsorbate search—which the Summary already flags as future work. That risk does not falsify the claim inside the stated scope, so it does not warrant moving from ACCEPT to CONDITIONAL. An honest non-finding is therefore appropriate: no load-bearing internal weakness was identified that would change the verdict.","tokens_in":16607,"tokens_out":537,"duration_ms":11026,"concrete_test":"Recompute the (100) and (¯201) surface free-energy diagrams of Fig. A.2 after adding one plausible larger-cell Ga-adlayer or dimer reconstruction (e.g., 2×1 or √2×√2 Ga-rich cells already noted as verified for stability of the bulk cuts) and re-evaluate γ(Δμ_O); if any such structure crosses below the stoichiometric line for Δμ_O > −2.5 eV, the claimed onset of Ga-rich stability would shift inside the ordinary growth window.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is supported inside the paper’s stated model: bulk-truncated symmetric slabs, G ≈ E_DFT (plus harmonic F_vib only for stoichiometric cases), and chemical-potential bounds from bulk stability. The reader’s weakest assumption correctly flags that larger-periodicity reconstructions or adsorbates are out of scope and could in principle move non-stoichiometric boundaries, but that is an explicit limitation rather than an internal inconsistency. Within the computed set, stoichiometric terminations remain lowest from roughly Δμ_O ≈ −2.5 to −0.5 eV across all orientations (Fig. 4 and A.2), PBEsol/PBE0 ordering agrees (Fig. 2, Table I), vibrations stay <0.2 J/m² and do not reorder surfaces, and the Ga-adlayer-like terminations on (100)/(¯201) appear only below ≈−2.5 eV, consistent with the cited MOVPE reducing-condition observations. No hidden algebraic or thermodynamic error undermines the claim as formulated.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.5","summary":"This manuscript presents a systematic DFT study of all symmetrically inequivalent low-index surfaces of monoclinic β-Ga₂O₃—(010), (100), (001), (¯201), (110), (111), and (11¯1)—using PBEsol and PBE0(0.26) in FHI-aims. Surface free energies are obtained from standard ab initio thermodynamics (Eqs. 1–7), with optional harmonic vibrational corrections for stoichiometric terminations (Eq. 4). The authors report a consistent energetic ordering across functionals, with (100) lowest (~0.6 J/m²) and (010)/(110) highest (~1.6–1.7 J/m²), vibrations below 0.2 J/m² up to 1000 K that do not reorder surfaces, and a linear coordination model (Eq. 8, Table II) linking stability to under-coordinated O and Ga environments. Thermodynamic diagrams (Fig. 4, A.2) show stoichiometric terminations stable over nearly the full bulk-stability window in Δμ_O, with Ga-rich adlayer-like terminations on (100) and (¯201) only under highly reducing conditions and limited O-rich terminations under strong oxidation; Wulff shapes (Fig. 5) are dominated by (100).","tokens_in":16772,"tokens_out":815,"duration_ms":14567,"significance":"The work fills a clear gap by treating all low-index orientations of β-Ga₂O₃ on equal footing with both semi-local and hybrid functionals, vibrational free energies, and full chemical-potential dependence. The consistent ordering with literature (Table I), the explicit prediction of Ga-adlayer-like terminations under reducing conditions that aligns with cited MOVPE observations, and the openly archived NOMAD data set are concrete strengths. The coordination model, while secondary and fitted, offers a transparent structural rationale that could guide higher-index screening. Within the stated scope of bulk-truncated terminations, the results provide a useful reference for growth and substrate selection in power electronics.","major_comments":[],"minor_comments":[{"comment":"In Sec. IV C and Fig. 4 the O-rich transition is written once as “ΔΔμ_O > −0.5 eV”; this is a typographical slip for Δμ_O.","section":"Sec. IV C / Fig. 4"},{"comment":"Table I header lists (201) without the overbar used elsewhere for (¯201); the same orientation should be labeled consistently.","section":"Table I"},{"comment":"Fig. A.1 atom-index plots would be easier to read if the surface-normal Δz panels used a common vertical scale or if the most displaced atoms were annotated.","section":"Appendix Fig. A.1"},{"comment":"The coordination model (Eq. 8) is fitted only to stoichiometric terminations; a brief remark on whether the same coefficients remain meaningful for the non-stoichiometric cases in Fig. A.2 would clarify its intended range of use.","section":"Sec. IV A, Eq. 8"},{"comment":"A short statement of slab thickness (or number of formula units) for each orientation would help readers assess residual finite-size effects beyond the force and vacuum criteria already given.","section":"Sec. III"}],"recommendation":"accept","confidential_remarks":"The manuscript is solid, incremental surface science rather than a conceptual breakthrough, but it is carefully executed and well matched to a specialized materials journal. I see no reason to delay it for reconstruction searches that the authors already flag as future work. Accept is appropriate."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"This is a competent, usable map of the seven symmetrically inequivalent low-index β-Ga2O3 surfaces. What is actually new is the breadth and integration: all orientations in one place, PBEsol vs PBE0(0.26), harmonic vibrations on the stoichiometric set, full Δμ_O diagrams, a simple coordination decomposition, and Wulff shapes—including the concrete prediction of Ga-adlayer-like terminations on (100) and (¯201) only under highly reducing conditions.\n\nThey do the standard things carefully. All-electron FHI-aims, tight/intermediate defaults, force threshold, k-point scaling, experimental O2 binding energy to cut functional bias on μ_O, and formation-enthalpy bounds are all spelled out. Ordering is consistent across functionals and with the literature table. Vibrations stay below ~0.2 J/m^{2} to 1000 K and do not reorder the faces. Stoichiometric terminations sit lowest from roughly −2.5 to −0.5 eV across the set; non-stoichiometric ones only at the extremes. Data are deposited in NOMAD. That is real, reproducible work in an active materials niche.\n\nThe soft spot is exactly the one they flag: bulk-truncated symmetric slabs only. Larger-periodicity reconstructions and adsorbates (H, water, etc.) are out of scope. If those lower non-stoichiometric energies inside the normal growth window, the stoichiometric-dominance window and the Ga-adlayer onset can move. That is a genuine limitation for growth interpretation, not an internal contradiction. The coordination model is a post-hoc linear fit; useful intuition, not a prediction engine. Absolute energies still scatter versus prior HSE/B3LYP numbers, as usual.\n\nWho it is for: people doing MOVPE/MBE, substrate choice, or morphology on β-Ga2O3. Not a methods paper and not a field-wide reset. I would cite the diagrams and the adlayer prediction when I need a reference surface-energy set. It deserves a serious referee, not a desk reject. Accept and move on; the residual risk is scope, not soundness inside the stated model.","headline":"Solid, useful unification of β-Ga2O3 low-index surface thermodynamics; claim holds inside a clearly scoped bulk-truncated model, with the main limit being reconstructions left out of scope.","tokens_in":17475,"tokens_out":573,"would_cite":true,"duration_ms":11890,"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":"Stoichiometric β-Ga₂O₃ surfaces stay lowest-energy across ordinary growth conditions; Ga adlayers appear only when oxygen is scarce.","keywords":["β-Ga2O3","surface free energy","stoichiometric terminations","Ga adlayer","oxygen chemical potential","Wulff construction","density-functional theory","coordination model"],"falsifier":"Surface-sensitive growth or spectroscopy experiments that detect a Ga adlayer (or its absence) on (100) or (−201) β-Ga₂O₃ precisely when the oxygen chemical potential crosses roughly −2.5 eV would confirm or refute the predicted transition.","tokens_in":17383,"feed_emoji":"🔬","tokens_out":945,"duration_ms":21087,"temperature":0.7,"pith_summary":"This paper maps the thermodynamic stability of every symmetrically distinct low-index surface of monoclinic β-Ga₂O₃ under the oxygen chemical potentials that matter for crystal growth. Using density-functional theory, the authors show that terminations keeping the bulk 2:3 Ga:O ratio win over nearly the whole stability window of the oxide, while non-stoichiometric cuts appear only at extreme reducing or oxidizing limits. Under highly reducing conditions they predict stable Ga-rich terminations that look like gallium adlayers on the (100) and (−201) faces—the orientations used most in epitaxy. A simple coordination model explains the ranking: under-coordinated oxygen and tetrahedral gallium raise the energy, whereas exposed under-coordinated octahedral gallium marks a stable face. Vibrational free-energy corrections stay small and do not reorder the surfaces. The result gives growers a concrete map of which faces and terminations to expect as oxygen pressure and temperature change.","feed_headline":"Ga₂O₃ surfaces stay stoichiometric until oxygen runs out","feed_subtitle":"DFT map shows Ga adlayers only under extreme reducing growth; (100) still dominates the crystal shape","key_machinery":"Ab initio surface free energy γ(T, p_O) from symmetric bulk-truncated slabs, with Ga and O chemical potentials linked by bulk equilibrium, plus a linear coordination model that expresses γ from the areal densities of under-coordinated tetrahedral Ga, octahedral Ga, and oxygen.","core_discovery":"Stoichiometric terminations of the low-index β-Ga₂O₃ surfaces are thermodynamically preferred over nearly the entire oxygen-chemical-potential range relevant to bulk stability; stable Ga-rich terminations resembling Ga adlayers form on (100) and (−201) only under highly reducing conditions (Δμ_O < −2.5 eV), while O-rich terminations appear only under strongly oxidizing conditions.","pith_inferences":["If intentional Ga-adlayer growth on (100) proves controllable, the same reducing window may be usable on (−201) substrates to engineer similar step-flow behavior.","Because under-coordinated tetrahedral Ga is strongly penalizing, any future reconstruction search should prioritize motifs that restore fourfold Ga2 coordination.","The small vibrational corrections suggest that finite-temperature morphology maps for this oxide can often omit phonons unless two terminations lie within ~0.1 J/m²."],"forward_implications":["Under ordinary MOVPE or melt-growth oxygen pressures, growers should expect stoichiometric (100), (−201), and (001) faces rather than reconstructed non-stoichiometric ones.","Highly reducing growth recipes that deliberately enter the Ga-adlayer regime on (100) can lower step-edge barriers and favor step-flow morphology.","Wulff shapes remain (100)-dominated (~42 % area) across the chemical-potential window, with (11−1), (001), and (−201) sharing most of the rest.","The coordination model supplies a quick filter for ranking higher-index stoichiometric orientations before full DFT.","Hybrid and semi-local functionals give the same energetic order, so cheaper PBEsol rankings are sufficient for morphology work."],"fun_headline_variants":["β-Ga₂O₃ surfaces stay stoichiometric across nearly all growth conditions","Ga adlayers form on (100) and (−201) only when oxygen is extremely scarce","Stoichiometric terminations dominate β-Ga₂O₃; non-stoichiometric need extremes","Under-coordinated O and tetrahedral Ga destabilize β-Ga₂O₃ surfaces","Vibrational terms stay below 0.2 J/m²; energetic order holds to 1000 K"],"cache_read_input_tokens":128,"weakest_assumption_plain":"The stability ranking rests on bulk-truncated slab cuts alone; larger reconstructions or adsorbate-covered surfaces are left out and could lower non-stoichiometric energies inside the normal growth window.","fun_headline_variants_meta":{"raw":{"variants":["β-Ga₂O₃ surfaces stay stoichiometric across nearly all growth conditions","Ga adlayers form on (100) and (−201) only when oxygen is extremely scarce","Stoichiometric terminations dominate β-Ga₂O₃; non-stoichiometric need extremes","Under-coordinated O and tetrahedral Ga destabilize β-Ga₂O₃ surfaces","Vibrational terms stay below 0.2 J/m²; energetic order holds to 1000 K"]},"model":"grok-4.5","effort":"low","cost_usd":0.004128,"raw_usage":{"total_tokens":1279,"prompt_tokens":832,"num_sources_used":0,"completion_tokens":102,"cost_in_usd_ticks":41284000,"prompt_tokens_details":{"text_tokens":832,"audio_tokens":0,"image_tokens":0,"cached_tokens":128},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":345,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":832,"tokens_out":102,"duration_ms":6433,"temperature":1.0,"reasoning_tokens":345,"cache_read_input_tokens":128,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-31T01:19:16.545284+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"Surface-sensitive growth or spectroscopy experiments that detect a Ga adlayer (or its absence) on (100) or (−201) β-Ga₂O₃ precisely when the oxygen chemical potential crosses roughly −2.5 eV would confirm or refute the predicted transition.","supporting_citations":[],"review_version":1}