{"id":"3a58a40b-b9a2-4013-aac1-24fd86f5dbb3","arxiv_id":"2607.18096","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Combining the variational polaron equations with the simplified string method gives optimized minimum-energy paths and adiabatic hopping mobilities without explicit supercells, validated on LiF and rutile TiO2.","lead":"The authors demonstrate a supercell-free method for computing the optimal hopping path of polarons—localized charges that drag a lattice distortion—using the string method on a variational polaron energy surface. The method yields hopping barriers, rates, and mobilities for LiF and rutile TiO2, matching measured electron mobility in TiO2 along the [001] direction.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"String method uses an unphysical metric: B-space gradients and arc length omit the 1/ω mass weighting, so claimed optimized MEPs may not be physical transition paths.","rationale":"Reader's conditional verdict is appropriate, but for a different reason. The reader focused on discarding delocalized transition states for [110]/[111] and mesh convergence. That is a real concern, yet the [001] agreement with experiment is not affected by it. The larger issue is that the string method's coordinate metric is inconsistent with the physical lattice dynamics. Since the central claim is that optimized MEPs provide transition states, this metric problem directly threatens the validity of all reported barriers and mobilities. It is testable and fixable, so the verdict remains conditional rather than reject; however, the condition should include a metric-corrected recomputation of at least one MEP before accepting the central quantitative claim.","tokens_in":21816,"tokens_out":11285,"duration_ms":490826,"concrete_test":"Recompute the TiO2 [001] and [110] MEPs (and the LiF hole [110] hopping MEP) using the physical metric: replace Eq. (10) by \\dot B_{qν} = -ω_{qν} ∇_{B_{qν}} E and reparametrize with arc length s_i = Σ_j (Σ_{qν} |B_{qν}^j - B_{qν}^{j-1}|² / ω_{qν})^{1/2} instead of Eq. (13). Compare ΔE_a and TS densities with Figs. 6–7. If ΔE_a shifts by more than ~20% (e.g., [001] rises from 3.86 meV to >10 meV), the central claim is not robust.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that the string method yields optimized MEPs and transition states for adiabatic polaron hopping. The implementation (Secs. IV.B–IV.C) evolves images with Eq. (10), \\dot B_i = -\\nabla_B E[A_i,B_i], and reparametrizes by equal L2 arc length in B (Eqs. 13–14). But B is not the physical coordinate. With Eq. (3), the mass-weighted normal coordinate is Q_{qν} ∝ B_{qν}/√ω_{qν}; the lattice kinetic energy is T = 1/2 Σ |\\dot Q|² = 1/2 Σ |\\dot B|²/ω_{qν}. Thus the physical steepest-descent force in B coordinates is -ω_{qν} ∂E/∂B_{qν}, and the proper arc length is Σ |δB_{qν}|²/ω_{qν}, not ∥δB∥. Using unweighted ∇_B E (Eq. 5) and L2 reparametrization defines a different curve; it is not the intrinsic reaction coordinate. Activation energies and transition-state geometries extracted from these curves (Figs. 6–7, Table I) are therefore not guaranteed to be the physical barriers. This is more load-bearing than the delocalized-TS exclusion flagged by the Reader, because if the metric is wrong all MEPs—including the [001] channel that supports the experimental agreement—are suspect. For low-frequency modes the 1/ω factor can be large, so the error need not be small.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript introduces a supercell-free approach to polaron hopping transport by combining the variational ab initio polaron equations with the simplified string method. The string method is used to optimize paths in the space of phonon coefficients B between self-trapped polaron states; along the path, the electronic coefficients are relaxed, and the resulting minimum-energy paths (MEPs) are used to extract activation energies, attempt frequencies, adiabatic transfer rates, and hopping mobilities. Applications to LiF (hole and electron polarons) and rutile TiO2 (electron polaron) are presented. The authors report multi-step hopping in LiF, near-barrierless rotation of the LiF hole polaron, anisotropic MEPs in TiO2, and a [001] electron mobility in rutile that agrees with experiment, while iBTE overestimates it.","tokens_in":22228,"tokens_out":14775,"duration_ms":156161,"significance":"The methodological proposal is significant and timely: it extends a variational reciprocal-space polaron formalism to transition-state search, avoiding supercell artifacts and providing explicit gradients. The paper includes code availability in ABINIT, materials-cloud data, and comparisons with MLMD, DMC, and experiment. If the transition paths are valid, the approach offers a scalable route to polaron-hopping mobility. However, two issues are load-bearing: the string method is implemented in B space with an unphysical Euclidean metric, and the [110]/[111] TiO2 results depend on discarding delocalized transition states without demonstrated convergence. For these reasons, the central quantitative claims are not yet fully supported.","major_comments":[{"comment":"The string method is formulated and implemented in B space using the Euclidean norm for both the force evolution and the reparametrization. From Eq. (3), the physical mass-weighted normal coordinate is Q_{qν} ∝ B_{qν}/√ω_{qν}; the lattice kinetic energy is T = 1/2 Σ|Q̇|² = 1/2 Σ|Ḃ|²/ω_{qν}. The intrinsic reaction coordinate therefore uses the metric Σ|δB|²/ω_{qν}, not Σ|δB|². Eq. (10) should be written in mass-weighted coordinates, and the arc length in Eq. (13) should be correspondingly weighted. As written, the optimized curve is not the minimum-energy path on the physical pPES; intermediate images and multi-step mechanisms (e.g., Fig. 2b) are not guaranteed to have physical meaning. A converged maximum of the curve would still be a stationary point of E_pol, but the paper does not verify this, and the metric used may direct the string to a different saddle. This concern applies to al","section":"IV.B–IV.C, Eqs. (3), (10), (13)–(14)"},{"comment":"For the [110] and [111] channels in rutile TiO2, the localized MEPs are reported only for n=9 and n=11, respectively; smaller BvK meshes yield delocalized transition states and are discarded as outside adiabatic hopping theory. No systematic convergence of the localized MEP with mesh size is shown, and for [111] the parameters are taken from a single mesh without thermodynamic-limit extrapolation. Because the anisotropic mobility comparison in Fig. 8 and the room-temperature values in Table I use these channels, the exclusion of delocalized paths is load-bearing. The authors should (i) define a quantitative localization criterion for the transition state, (ii) report the energy and barrier of the delocalized paths (they appear in Fig. 6 but their barriers are not given), and (iii) test convergence of the localized MEP with increasing n, or explain why the n=9/n=11 results are converged.","section":"II.C, Figs. 6–7, Table I"},{"comment":"Activation energies are extrapolated to the thermodynamic limit by linear fits in N_p^{-1/3}; no uncertainty estimates or justification for the linear dependence are provided. For the [001] channel the extrapolated barrier is 3.86 meV, of the order of the scatter expected from such fits and much less than k_BT at room temperature. The agreement of the [001] mobility with experiment could therefore be sensitive to the extrapolation procedure. Please provide error bars or a more detailed convergence analysis, and clarify how many mesh points enter each extrapolation.","section":"II.C, Fig. 7"}],"minor_comments":[{"comment":"The number of images N and the dependence of the converged barrier on N are not reported; please provide these convergence details for reproducibility.","section":"IV.C"},{"comment":"The sentence 'The initial linearly-interpolated energy barriers, agree' contains a grammatical error and should be rewritten. Also, the text should clarify which polaron-density isosurface is used for the delocalized transition states in Fig. 6.","section":"II.C"},{"comment":"The Materials Cloud Archive entry is marked 'draft'; a stable DOI should be provided in the published version.","section":"Data availability"},{"comment":"The reorganization energy and electronic coupling are defined but not used in the reported rates and mobilities; the role of these quantities in the adiabatic regime should be stated explicitly, or their reporting should be justified.","section":"IV.E, Eqs. (19)–(22)"}],"recommendation":"major_revision","confidential_remarks":"The main obstacle is the unphysical metric used in the string method; it is fixable in principle but may require redoing all MEP calculations. The delocalized-transition-state exclusion also needs stronger numerical support. If the authors can show that the B-Euclidean string method converges to true stationary points and that the barriers are robust, the paper would be a strong candidate for publication. The paper builds appropriately on the authors' previous formalism; this is not a novelty concern."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Worth a look: this is the first time the simplified string method has been brought to the variational polaron equations, and it gives a clean path to adiabatic hopping barriers and mobilities without supercells. The best evidence is the [001] electron-polaron channel in rutile TiO2: the optimized barrier drops from 12 to 4 meV, and the resulting mobility is in the right ballpark against experiment, while iBTE badly overshoots. The LiF hole-polaron picture with rotations and a two-centre X2- transition state is a nice demonstration that linear interpolation can be badly misleading. The authors also ship the implementation in ABINIT and put data on Materials Cloud, which is the right way to do this.\n\nThe stress-test about the B-space metric is real but probably not fatal: the physical mass-weighted coordinate is Q proportional to B/sqrt(omega), so the gradient flow and arc length should carry the 1/omega weight, meaning the converged curve in the unweighted B-metric is not the physical IRC. However, the maximum along that curve still has zero gradient, so the barrier height from a converged string should still be the energy of a true stationary point, assuming the algorithm lands on the intended saddle. What the paper does not show is that the stationary point is the correct first-order saddle — a Hessian check or a direct saddle search would settle it. That needs to be addressed, but it does not by itself sink the central claim.\n\nThe bigger soft spot is the treatment of TiO2 [110] and [111]. For those channels the authors simply discard the MEPs that show delocalized transition states, on the grounds that they fall outside adiabatic hopping theory. That is a physical judgment call, but the paper picks the largest mesh that still gives a localized TS and never shows convergence from above; n=9 and n=11 might just be meshes that happen to suppress delocalization. The anisotropic mobilities in Table I depend on this. It is the weakest load-bearing section.\n\nMinor issues: the attempt frequency nu_eff is an average over the initial-state phonon cloud, and the thermodynamic-limit extrapolation is linear through a few points, so neither should be pushed too hard. Self-citation is background, not a problem here.\n\nVerdict: send it to a serious referee. The method and the [001] result are worth peer review, but the [110]/[111] analysis and the metric justification need real attention. I would want it revised rather than accepted as is.","headline":"A genuinely new string-method-plus-variational-polaron combination with a promising TiO2 [001] result; the [110]/[111] channels and the unwritten metric assumption both need referee attention.","tokens_in":22676,"tokens_out":6437,"would_cite":true,"duration_ms":72053,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper shows that polaron hopping barriers, transition states, and adiabatic mobilities can be computed without supercells by combining variational polaron equations with the string method, and that the resulting rutile TiO2 electron mo","keywords":["polaron hopping","small polaron","minimum-energy path","string method","variational polaron equations","rutile TiO2","LiF","adiabatic transport"],"falsifier":"Compute the [110] and [111] minimum-energy paths on k/q meshes denser than 11×11×22 (or include the delocalized transition paths shown in the paper in the transition-state rate) and check whether activation energies of about 81 meV and 78 meV and the mobilities remain stable; alternatively, measure direction-resolved electron mobility in oriented rutile TiO2 single crystals between 200 and 400 K and compare with the predicted [001]/[110]/[111] anisotropy.","tokens_in":21737,"feed_emoji":"⚡","tokens_out":4465,"duration_ms":41949,"temperature":0.7,"pith_summary":"Polaron hopping controls charge transport in many materials where electrons get trapped by their own lattice distortion, but standard supercell DFT calculations are expensive and unreliable for such transitions. This paper introduces a method that skips supercells entirely: it works in reciprocal space with a variational polaron energy, and uses the simplified string method to find the minimum-energy path between two polaron states. The path yields the transition state, the activation barrier, an effective attempt frequency, and through transition-state theory, the adiabatic hopping rate and mobility. Applied to rutile TiO2, the computed electron-polaron mobility along the most favourable [001] direction matches measured mobilities, while the Boltzmann transport equation overestimates them because it assumes delocalized band carriers. In LiF, the method uncovers rotation and multi-step hopping mechanisms for the hole polaron.","feed_headline":"Supercell-free method reproduces TiO2 polaron hopping mobility","feed_subtitle":"Variational paths in reciprocal space match measured electron mobility where band transport overshoots.","key_machinery":"The central object is the variational polaron energy functional Epol[A,B] depending on electronic coefficients A and phonon deformation coefficients B, whose gradients provide forces on the polaron potential energy surface. The simplified string method evolves a discretized path between two polaron configurations using only the phonon gradient, with the electronic coefficients re-optimized for each image at fixed deformation; equal-arc-length reparametrization keeps the path smooth. This gives minimum-energy paths and transition states without real-space supercells.","core_discovery":"The central claim is that the minimum-energy paths for polaron transfer, and therefore the activated barriers and hopping mobilities, can be obtained variationally in reciprocal space without constructing supercells. The authors combine the ab initio polaron equations with the simplified string method, representing the path by phonon deformation fields that determine the optimal electronic polaron via constrained minimization. This yields optimized adiabatic transition states with lower barriers than linear interpolation: for rutile TiO2 the [001] barrier drops from 12 meV to 4 meV, and the resulting room-temperature mobility of 0.446 cm2/V·s agrees with experiment, whereas iterative Boltzma","pith_inferences":["Because the framework avoids supercells and uses reciprocal-space gradients, it could be extended to all-coupling variational polaron theories, potentially covering weak- and intermediate-coupling regimes where the present strong-coupling adiabatic approximation is less reliable.","The reported [110] and [111] mobilities rest on discarding transition states where the polaron delocalizes across the simulation cell; a natural test is to include those delocalized paths in a rate theory or compare with diagrammatic Monte Carlo at lower temperatures.","The crossover temperature T* = ΔEa/kB is about 45 K for the TiO2 [001] channel, predicting a measurable negative-to-positive temperature dependence of mobility that oriented single-crystal experiments could confirm.","The same variational string method could be applied to hole polarons, exciton polarons, or defect-bound polarons in other oxides and halides, where the formalism only needs the electronic eigenvalues, phonon dispersions, and electron-phonon matrix elements."],"forward_implications":["In rutile TiO2 the adiabatic hopping mobility computed from optimized paths agrees with experiment, while band-like Boltzmann transport substantially overestimates it, indicating that self-trapped carriers demand a hopping description.","Optimizing the energy path matters: linear interpolation overestimates activation barriers (e.g., TiO2 [001] from 12 meV to 4 meV) and would underestimate mobilities.","The method resolves anisotropic and multi-step hopping mechanisms: for the LiF hole polaron, inter-site translation through a two-centre transition state (barrier about 438 meV) is rate-limiting, while on-site rotation is almost barrier-less (20 meV).","The approach yields direction-resolved mobilities, identifying [001] as the dominant electron-polaron channel in rutile TiO2, with [110] and [111] channels requiring larger simulation meshes to remain localized.","The effective attempt frequency can be computed as a phonon-weighted average at the initial polaron configuration, rather than resorting to a single longitudinal-optical mode."],"fun_headline_variants":["Supercell-free method matches measured TiO2 polaron mobility","Polaron hopping barriers without supercells, validated on TiO2","Variational paths in reciprocal space beat band transport for TiO2","New approach reproduces polaron mobility in TiO2 without supercells","Polaron hopping via reciprocal space: TiO2 mobility matches experiment"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The load-bearing premise is that a polaron that spreads out across the whole simulation cell at the transition state is not a valid adiabatic hopping event and can be discarded; for the [110] and [111] directions in rutile TiO2, the reported anisotropic mobilities change if that premise fails, and the convergence of the localized path with cell size is not fully demonstrated beyond the two largest meshes used.","fun_headline_variants_meta":{"raw":{"variants":["Supercell-free method matches measured TiO2 polaron mobility","Polaron hopping barriers without supercells, validated on TiO2","Variational paths in reciprocal space beat band transport for TiO2","New approach reproduces polaron mobility in TiO2 without supercells","Polaron hopping via reciprocal space: TiO2 mobility matches experiment"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00019,"raw_usage":{"total_tokens":1154,"prompt_tokens":697,"completion_tokens":457,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":441,"completion_tokens_details":{"reasoning_tokens":383}},"tokens_in":441,"tokens_out":457,"duration_ms":5069,"temperature":1.0,"reasoning_tokens":383,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T16:00:26.151792+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the [110] and [111] minimum-energy paths on k/q meshes denser than 11×11×22 (or include the delocalized transition paths shown in the paper in the transition-state rate) and check whether activation energies of about 81 meV and 78 meV and the mobilities remain stable; alternatively, measure direction-resolved electron mobility in oriented rutile TiO2 single crystals between 200 and 400 K and compare with the predicted [001]/[110]/[111] anisotropy.","supporting_citations":[],"review_version":1}