{"id":"e3650070-a089-4976-8985-7d9c1cea1696","arxiv_id":"2607.08860","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":5.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"Rutile TiO2 has a clear ferroelastic path to the CaCl2-type (Pnnm) phase at ~13.5 GPa, but that phase is thermodynamically suppressed by more stable higher-coordination polymorphs.","lead":"Rutile TiO2 develops a ferroelastic instability toward the CaCl2-type phase near 13.5 GPa via elastic softening and a soft B1g phonon, yet enthalpy shows this phase is only weakly metastable and overtaken by columbite and baddeleyite. The work explains why the intermediate is almost never seen experimentally despite being the natural dynamical pathway.","discovery_kind":"extension","skeptic_critique":{"model":"grok-4.5","headline":"No significant objection identified beyond the functional-sensitivity caveat already flagged by the reader.","rationale":"The multi-diagnostic evidence (elastic Born criterion, soft B1g, double-well strain, Landau scaling of δ^{2}) cleanly establishes the ferroelastic pathway, while the Birch–Murnaghan enthalpy curves show why that pathway is bypassed. The only soft spot is the meV-scale ranking under mixed functionals, already noted by the reader as typical rather than fatal. No additional load-bearing flaw (symmetry misassignment, missing LO–TO treatment, incomplete phase set, etc.) appears. Therefore the ACCEPT verdict stands; the concrete functional-consistency check is a useful but non-decisive verification step.","tokens_in":13310,"tokens_out":455,"duration_ms":5268,"concrete_test":"Recompute the relative enthalpies of rutile, Pnnm, columbite and baddeleyite at 15–25 GPa with a single consistent functional (e.g., PBEsol or a hybrid) and identical all-electron settings; if the Pnnm–rutile gap remains ≲15 meV/f.u. while columbite/baddeleyite stay lower by hundreds of meV, the suppression claim is robust.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper’s central claim is that a clear ferroelastic pathway (C11−C12 collapse, B1g condensation, double-well strain landscape at ~13.5 GPa) exists yet is thermodynamically suppressed by columbite and baddeleyite, leaving Pnnm only weakly metastable (~7 meV/f.u. vs rutile at 20 GPa). The elastic, phonon and Landau analyses are mutually consistent and the enthalpy ranking is transparent. The only load-bearing uncertainty is the quantitative reliability of those few-meV enthalpy differences under PBE (elastic/enthalpy) versus LDA (phonons), which the reader already identified. No deeper internal inconsistency or missing control is apparent; the functional issue is standard for the field and does not overturn the qualitative conclusion that Pnnm is not the ground-state intermediate.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.5","summary":"The manuscript investigates the long-elusive CaCl2-type (Pnnm) intermediate in the high-pressure sequence of rutile TiO2. Using all-electron APW+lo DFT (WIEN2k/PBE) for elastic constants and enthalpies, together with DFPT phonons (Quantum ESPRESSO/LDA), the authors show that rutile develops a ferroelastic instability near 13.5 GPa: the combination C11-C12 softens and violates the generalized Born criterion, a double-well energy landscape appears with respect to orthorhombic strain δ, and the B1g phonon condenses, producing coordinated TiO6 rotations that lower the symmetry to Pnnm. Landau analysis of δ^{2}(P) confirms a continuous transition. Enthalpy comparisons (Birch–Murnaghan EOS) nevertheless demonstrate that Pnnm is only weakly stabilized relative to rutile (~7 meV/f.u. at 20 GPa) and remains far above columbite and baddeleyite. The CaCl2-type phase is therefore identified as the intrinsic ferroelastic response of rutile that is thermodynamically suppressed, explaining its experimental elusiveness.","tokens_in":13505,"tokens_out":1150,"duration_ms":11439,"significance":"The work supplies a coherent, multi-probe (elastic + dynamical + thermodynamic) resolution of a long-standing puzzle in the high-pressure polymorphism of TiO2 and, by extension, of rutile-type dioxides. The mutual consistency of the C11-C12 collapse, B1g softening, double-well strain landscape, and Landau scaling is a clear strength; the all-electron elastic constants match experiment well and the phonon frequencies reproduce the known Raman soft mode. The central claim—that a well-defined ferroelastic pathway exists yet is over-ridden by higher-coordination phases—is falsifiable and of direct interest to the high-pressure materials community. The quantitative reliability of the few-meV enthalpy differences remains the principal caveat, but the qualitative ranking is robust enough to constitute a useful advance.","major_comments":[{"comment":"Methodology §2.1–2.3 and Results enthalpy discussion (Fig. 5): elastic constants and enthalpies are obtained with PBE while phonons are computed with LDA. The decisive energy scale that places Pnnm only ~7 meV/f.u. below rutile (and hundreds of meV above baddeleyite) is comparable to typical residual DFT errors and to known PBE/LDA volume and soft-mode discrepancies for TiO2. A single-functional cross-check (or at least an explicit statement of how the ranking survives a volume-corrected or hybrid calculation) is needed to underwrite the claim that Pnnm is thermodynamically suppressed rather than merely an artifact of functional choice.","section":null},{"comment":"Results, comparison with Liu et al. (elastic stability threshold ~15 GPa vs their 19.2 GPa): the discrepancy is attributed to differences in C11 under pressure, yet no additional convergence tests (RMTKmax, k-mesh, or strain-amplitude dependence of the IRELAST fits) are reported for the high-pressure elastic constants themselves. Because the critical pressure enters both the Born criterion and the Landau Pc, a short convergence table or supplementary plot would strengthen confidence that the 13.5 GPa value is not basis-set limited.","section":null}],"minor_comments":[{"comment":"Table 1: experimental ranges are given as single values with letter superscripts; a clearer presentation of the experimental spread (or explicit citation of each datum) would help the reader judge the 3–12 % deviations.","section":null},{"comment":"Fig. 2 caption: the energy offsets applied for visual clarity (0.005, 0.021, 0.041 Ry) should also be stated in the main text or in a table so that absolute well depths can be recovered without digitizing the figure.","section":null},{"comment":"Phonon section: the choice of LDA is justified by known PBE/pseudopotential issues for rutile, but a brief remark on whether the B1g Grüneisen parameter remains negative under PBE (or a hybrid) would close the functional loop more cleanly.","section":null},{"comment":"References: a few recent experimental high-pressure Raman/XRD studies on TiO2 that post-date Montanari & Harrison (2004) could be added for completeness, even if they do not report Pnnm.","section":null},{"comment":"Typographical: “CaCl2-type” is occasionally written without the subscript formatting consistency; “P42/mnm” vs “P4_2/mnm” should be uniform.","section":null}],"recommendation":"minor_revision","confidential_remarks":"The manuscript is solid and the central narrative is persuasive. The functional-sensitivity issue is standard for the field and does not warrant major revision or rejection; a short additional calculation or a carefully worded caveat will suffice. Fit for a solid-state / high-pressure materials journal is good."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"This paper does one useful thing cleanly: it shows that rutile TiO2 really does have the classic ferroelastic route to CaCl2-type (C11–C12 collapse, B1g condensation, double-well orthorhombic strain at ~13.5 GPa) and then demonstrates why that phase is almost never seen—its enthalpy gain versus rutile is only a few meV while columbite and baddeleyite sit hundreds of meV lower. That ranking, not the soft-mode story itself, is the new piece.\n\nMontanari & Harrison already had the elastic and B1g softening near 13 GPa. What is added here is the explicit Landau analysis of the strain double-well, full pressure-dependent phonon dispersions that confirm the soft mode, and especially the consistent Birch–Murnaghan enthalpy curves that put Pnnm in its place relative to the two competing phases. The all-electron elastic constants match experiment well; the LDA phonons recover the Raman frequencies and the known soft B1g; the linear δ^{2}(P–Pc) plot is textbook. The calculations are transparent and free of free parameters.\n\nThe soft spot is the usual one for this energy scale: PBE for elastic/enthalpy work and LDA for phonons, with differences of only ~7 meV/f.u. at 20 GPa. That is comparable to residual DFT error, so the precise numbers should not be over-read. But the qualitative conclusion—that Pnnm is only weakly metastable and is thermodynamically suppressed—does not hinge on the last meV; the volume collapse of the higher-coordination phases supplies a large PV advantage that is robust. No deeper inconsistency shows up.\n\nThis is for people who work on high-pressure dioxides or ferroelastic transitions in rutile-type oxides. It supplies a transferable diagnostic (elastic + phonon + enthalpy) and closes a long-standing experimental puzzle without overclaiming. I would send it to peer review; a referee can ask for a short functional-sensitivity note, but the central argument already stands. Worth reading and, for anyone in the area, worth citing.","headline":"Clean DFT unification of the long-noted rutile ferroelastic pathway with the thermodynamic reason Pnnm never sticks around; the meV-scale enthalpy ranking is the real addition.","tokens_in":14103,"tokens_out":536,"would_cite":true,"duration_ms":6298,"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":"Rutile TiO2 has a clear ferroelastic path to the CaCl2-type phase near 13.5 GPa, but thermodynamics keeps that phase only weakly metastable and experimentally elusive.","keywords":["CaCl2-type","high-pressure phases","ferroelastic transition","rutile TiO2","elastic instability","B1g phonon","phase competition"],"falsifier":"A high-resolution static-compression experiment that isolates a pure rutile sample, applies truly hydrostatic pressure through the 12–20 GPa window, and either detects (or definitively rules out) the orthorhombic strain and oxygen displacements of Pnnm before columbite or baddeleyite appear would decide the claim.","tokens_in":14231,"feed_emoji":"⚡","tokens_out":976,"duration_ms":14883,"temperature":0.7,"pith_summary":"This paper asks why the CaCl2-type intermediate that appears in many other rutile oxides is almost never seen in high-pressure TiO2. All-electron elasticity and phonon calculations show that rutile itself becomes unstable to an orthorhombic shear: the C11-C12 modulus softens, a B1g oxygen-rotation mode condenses, and a double-well energy landscape develops at about 13.5 GPa, producing the Pnnm (CaCl2-type) structure. Enthalpy comparisons, however, place that phase only a few millielectronvolts below rutile and far above the competing columbite and baddeleyite phases. The authors therefore conclude that the CaCl2-type structure is the lattice’s intrinsic ferroelastic response, yet is thermodynamically suppressed and can appear only as a transient or weakly metastable intermediate—explaining its long-standing experimental elusiveness.","feed_headline":"TiO2's CaCl2 phase is real but thermodynamically smothered","feed_subtitle":"A ferroelastic soft mode opens the door near 13.5 GPa; denser rivals slam it shut.","key_machinery":"The ferroelastic order parameter—an orthorhombic strain δ coupled to the zone-center B1g oxygen-rotation mode—creates a double-well energy landscape whose condensation yields Pnnm; thermodynamic ranking is then decided by enthalpy H = E + PV against denser competitors.","core_discovery":"Rutile TiO2 undergoes a genuine ferroelastic instability at ~13.5 GPa driven by C11-C12 softening and B1g phonon condensation, which lowers the symmetry to the CaCl2-type (Pnnm) phase via coordinated TiO6 octahedral rotations; nevertheless the resulting phase is only weakly stabilized relative to rutile and remains higher in enthalpy than columbite and baddeleyite, so it never becomes a stable high-pressure polymorph.","pith_inferences":["If the small enthalpy gap is real, non-hydrostatic stress or rapid compression may be the only practical ways to populate the Pnnm well long enough for structural detection.","Functional choice that systematically over-stabilizes higher-coordination phases could artificially shrink the CaCl2-type window; a hybrid or meta-GGA re-ranking would test that bias.","The same ferroelastic soft-mode pathway may operate in doped or nanostructured TiO2 where surface or defect free energy could temporarily tip the enthalpy balance."],"forward_implications":["The CaCl2-type phase should be treated as a short-lived intermediate rather than a stable entry on the TiO2 phase diagram.","Direct rutile-to-baddeleyite (or rutile-to-columbite) paths remain the expected experimental routes under ordinary hydrostatic compression.","Raman or X-ray signatures of B1g softening and small orthorhombic strain may still be observable if kinetics temporarily trap the lattice in the ferroelastic well.","The same elasticity-plus-enthalpy filter can be applied to other rutile dioxides to predict which ones will display a stable CaCl2-type window."],"fun_headline_variants":["Rutile TiO2 softens into CaCl2-type at 13.5 GPa then loses to denser phases","Ferroelastic B1g path opens Pnnm in TiO2 but enthalpy keeps it buried","C11-C12 collapse creates weak CaCl2 intermediate that never stabilizes","TiO2's orthorhombic strain path is real yet thermodynamically smothered","CaCl2-type TiO2 forms via octahedral rotations then yields to columbite"],"cache_read_input_tokens":128,"weakest_assumption_plain":"The ranking of phases rests on enthalpy differences of only a few millielectronvolts per formula unit obtained with two different density functionals whose known volume and soft-mode errors for TiO2 are comparable to those tiny energy scales.","fun_headline_variants_meta":{"raw":{"variants":["Rutile TiO2 softens into CaCl2-type at 13.5 GPa then loses to denser phases","Ferroelastic B1g path opens Pnnm in TiO2 but enthalpy keeps it buried","C11-C12 collapse creates weak CaCl2 intermediate that never stabilizes","TiO2's orthorhombic strain path is real yet thermodynamically smothered","CaCl2-type TiO2 forms via octahedral rotations then yields to columbite"]},"model":"grok-4.5","effort":"low","cost_usd":0.003916,"raw_usage":{"total_tokens":1282,"prompt_tokens":848,"num_sources_used":0,"completion_tokens":108,"cost_in_usd_ticks":39160000,"prompt_tokens_details":{"text_tokens":848,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":326,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":848,"tokens_out":108,"duration_ms":4461,"temperature":1.0,"reasoning_tokens":326,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-13T06:12:05.580776+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"A high-resolution static-compression experiment that isolates a pure rutile sample, applies truly hydrostatic pressure through the 12–20 GPa window, and either detects (or definitively rules out) the orthorhombic strain and oxygen displacements of Pnnm before columbite or baddeleyite appear would decide the claim.","supporting_citations":[],"review_version":1}