{"id":"2bca6664-ac5b-49fc-a26c-d03f6232b1a9","arxiv_id":"2411.10839","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"Load stress during solid-state lithium deposition shifts surface energy enough to favor slow (001) grains; an amorphous lithium-silicon seed layer restores fast (101) grains and raises the critical current density.","lead":"Using a thermodynamic model and phase-field simulations, the authors explain why lithium deposited on copper in solid-state batteries grows with a slow (001)-oriented texture, and show that a thin amorphous silicon seed layer switches the texture to fast-growing (101) grains. The work links battery pressure and temperature to metal crystal orientation, and demonstrates a battery that avoids short-circuiting at higher charge rates.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The model relies on DFT surface energies at ~3% lattice strain, but the experiments use only 5 MPa stack pressure; Li's elastic strain is ~0.04%, so the load-stress-induced (001) selection mechanism is not connected to the stated boundary conditions.","rationale":"The reader's weakest assumption is the uncalibrated strain-mobility relation (Eq. S9) and the absence of sensitivity analysis. That is a legitimate concern, but it operates downstream of a more fundamental input: the magnitude of lattice strain under the experimental stack pressure. The central explanatory claim is that load stress deforms the Li lattice enough to change surface-energy anisotropy and thereby select (001) grains at 25°C. The experiments use 5 MPa, which for Li implies roughly 0.04% elastic strain rather than the 3% used in the DFT data. If the strain is that small, the surface-energy anisotropy is nearly the unstrained one, and the model itself would favor (101) grains even at room temperature. Thus the observed 25°C (001) texture, if reproducible, would not be explained by the proposed mechanism unless an additional source of large lattice strain is identified and quantified. I therefore move the verdict from CONDITIONAL to UNVERDICTED: the central mechanism is not established by the current boundary conditions. I would not reject outright because the EBSD observations and the empirical seed-layer improvement remain valuable regardless of the mechanism, and the proposed strain check could resolve the issue.","tokens_in":15278,"tokens_out":8840,"duration_ms":110037,"concrete_test":"Compute the actual lattice strain in Li for the 5 MPa stack pressure using its elastic constants (e.g., ε ≈ P/B = 3.6×10^-4 or uniaxial ε ≈ 5 MPa / 7.8 GPa ≈ 6×10^-4), then take the strain-dependent DFT surface energies from Fig. S2 at that strain and rerun the phase-field temperature analysis. If (001) no longer wins at 25°C, the load-stress mechanism is not supported by the stated experimental boundary condition; an independent check would be to measure residual lattice strain in the plated Li by XRD or EBSD-based methods and compare it with the 3% strain used in Fig. S2.","verdict_should_be":"UNVERDICTED","load_bearing_attack":"The central mechanism requires load stress to alter Li surface-energy anisotropy enough to favor (001) at 25°C. The supporting DFT data (Fig. S2, ref. 16) are reported at ~3% lattice strain, and the phase-field inputs are described as DFT-informed under load stress. However, the validating experiments are pellet cells with a 5 MPa stack pressure (Experimental Procedures), and even the commonly cited solid-state values are >10 MPa. For Li, the bulk modulus is ~14 GPa (Fig. 5A) and the yield strength is <1 MPa (LePage et al.). The elastic lattice strain at 5 MPa is therefore P/B ≈ 3.6×10^-4, i.e., ~0.04%, two orders of magnitude below 3%; plastic/creep deformation does not change lattice constants and so cannot produce the assumed surface-energy shift. Scaling the DFT anisotropy (~0.08 J/m² at 3% strain for (101)) linearly to 0.04% gives ~0.001 J/m², a surface-energy-density difference of only ~10^2 J/m³ at h = 10 μm, whereas Eq. S9 with σy = 0.55 MPa and εT−εc = 0.4 gives strain-energy differences of order 10^3–10^4 J/m³. Under realistic strains the model would then predict (101) even at 25°C, contradicting the claimed mechanism for the observed (001) texture. This also weakens the seed-layer rationale, which assumes bulk-modulus softening reduces Li lattice strain. The manuscript provides no strain measurement or stress-to-strain conversion to bridge this gap.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper proposes a thermodynamic framework and phase-field model for grain selection during electrodeposition of soft metals (Li, Na, K), in which grain growth is governed by competition between surface energy and an atomic-mobility-dependent intrinsic strain energy (Eq. 2/S9). The model is used to predict that under high stacking pressure in solid-state batteries, Li deposition at 25 °C selects (001)-oriented grains because strain-enhanced surface-energy anisotropy outweighs the low diffusion barrier of (101); at 80 °C, or on a soft amorphous Li0.7Si0.3 seed layer, (101) grains dominate. These predictions are compared with PFIB-EBSD texture measurements of Li deposited in anode-free solid-state cells at 25 °C and 80 °C, with a liquid-electrolyte control, and with rate-performance tests of Si-seeded cells. The EBSD data show the predicted temperature-dependent texture switch and 50% (101) texture with the Si seed layer at 25 °C.","tokens_in":15656,"tokens_out":11024,"duration_ms":105343,"significance":"The manuscript combines DFT-derived surface energies and diffusion barriers with a phase-field model and original cross-sectional EBSD measurements, a rare direct characterization of Li texture in solid-state cells. If the proposed mechanism is quantitatively correct, the work offers a design principle—soft amorphous seed layers with bulk modulus below 30 GPa—that is falsifiable and directly relevant to anode-free solid-state batteries. The experimental methodology (PFIB-EBSD at 7 kV with pattern matching) is a valuable contribution. The main caveats are the uncalibrated fitting constants in the intrinsic-strain and phase-field bridging equations and the missing stress-to-strain bridge between the 5 MPa stack pressure and the 3% lattice-strain DFT inputs; these issues prevent the paper's quantitative predictions from being taken at face value.","major_comments":[{"comment":"The intrinsic-strain relation (Eq. S9) contains an uncalibrated fitting constant β, and the authors set β = 1 with the statement that this 'does not affect the trend of this analysis.' Because β enters the exponent of the strain–mobility relation, the predicted temperature crossover and the phase map of Fig. 4 depend on β, as well as on the assumed yield strength (0.55 MPa), the assumed strain limits (±0.2), and the diffusion pre-factor D0. No sensitivity analysis is given for any of these parameters, so the quantitative predictions—including the location of the strain-energy/surface-energy boundary and the 25 °C versus 80 °C switch—are not secured by the derivation. The manuscript should either calibrate β and the strain limits against independent stress/strain or grain-growth measurements, or systematically show that the qualitative texture predictions are invariant over a plausible parameter range.","section":"Experimental Procedures, Eq. S9–S10 and 'Temperature analysis for grain selection growth'"},{"comment":"The proposed mechanism for the 25 °C (001) texture relies on load-stress-induced surface energy anisotropy computed by DFT at about 3% lattice strain (ref. 16, Fig. S2). The validating experiments, however, are pellet cells with a 5 MPa stack pressure. For Li, with a bulk modulus of about 14 GPa, 5 MPa produces an elastic lattice strain of only about 0.04%, and plastic/creep deformation does not change the lattice constants that enter the DFT surface-energy calculations. The manuscript does not provide a stress-to-strain conversion, a lattice strain measurement, or an argument that the 3% DFT condition is representative of the experimental conditions. At a linear scaling of the surface-energy anisotropy to 0.04% strain, the surface-energy density difference and the strain-energy difference of Eq. S10 would be orders of magnitude too small to favor (001) at 25 °C, contradicting the claimed mechanism. This gap also weakens the seed-layer rationale, which assumes that reducing the substrate bulk modulus lowers the Li lattice strain. The authors should supply experimental strain data or a quantitative mechanical model that connects the applied stack pressure to the lattice strain used in the DFT inputs.","section":"Results & Discussion, 'Load stress induced selective grain growth'; Figs. 3 and S2; Experimental Procedures"},{"comment":"The phase-field mobility and gradient-energy coefficients are assigned through exponential bridging laws, Eq. S14 and Eq. S15, with fitting constants (L1 = 7.5, L2 = 0.27, κ1 = 4.1×10^-9, κ2 = 20) that are chosen solely to keep the normalized values within 0.1–10. These constants are not calibrated against experimental grain-growth data, and the manuscript does not report the actual values of L_q and κ_q for each orientation or test how the simulated grain-proportion statistics (e.g., 57.1% (001) at 25 °C versus 48.7% (101) at 80 °C) depend on these choices. As a result, the quantitative texture fractions and the phase-map boundaries should be regarded as qualitative illustrations rather than parameter-free predictions.","section":"Experimental Procedures, Eqs. S14–S15"},{"comment":"The strain-energy density in Eq. S5 is written as F_strain = σ_y ε - (1/2)E ε_y^2, which is linear in the signed strain ε. With the assumed compressive strain ε_c = -0.2, this expression yields negative strain-energy densities for compressive grains, which is physically inconsistent with a plastic-work density of σ_y|ε|. The derivation of ΔF_strain in Eq. S10 should use the absolute value of the plastic strain (or otherwise justify the sign convention), because the magnitude and even the sign of the strain-energy difference between tensile and compressive grains can change under the current formulation. The authors should clarify this point and verify that the predicted texture selection is not an artifact of the signed-strain energy.","section":"Experimental Procedures, Eqs. S4–S7"}],"minor_comments":[{"comment":"The notation for the silicon seed layer is inconsistent: the abstract uses LixSi1-x (0.50<x<0.79), while the main text refers to Li0.7Si0.3 and Li3.75Si; please use one convention throughout.","section":"Abstract and text"},{"comment":"The phrase 'load stress-induced surface energy anisotropy' in the abstract and introduction is not directly tied to Eq. 2/S9, which describes intrinsic strain from atomic mobility; the load-stress connection enters only through the assumed surface-energy change. Please make this distinction explicit.","section":"Introduction and Eq. 2"},{"comment":"The statement 'Generally, ΔF_strain is positive and rises as temperature increases' is not obviously true from Eq. S10, since both exponential terms tend toward zero at sufficiently high temperature; please clarify the parameter regime or provide the data underlying Fig. 3A.","section":"Temperature effects section"},{"comment":"The description '75 mg of LPSCl was compressed at 370 MPa' does not specify the area of the pellet; please report the pellet diameter or area so the compaction pressure can be evaluated.","section":"Experimental Procedures, Full cells assembling"},{"comment":"For a modeling paper, depositing the phase-field code and fitting scripts in a public repository would improve reproducibility; the current statement only allows requests to the lead contact.","section":"Data and code availability"}],"recommendation":"major_revision","confidential_remarks":"The paper presents an appealing qualitative mechanism and valuable EBSD data, but the quantitative claims are not yet supported because of the uncalibrated parameters and the unbridged gap between the 5 MPa stack pressure and the 3% lattice-strain DFT inputs. The requested revisions are substantial but within scope: a sensitivity analysis and a mechanical strain estimate would be enough to make the central claims defensible."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my take. The paper is an interesting package: it shows by PFIB-EBSD that Li electrodeposited in solid-state cells has (001) texture at 25°C and (101)-preferred texture at 80°C, and that an amorphous LixSi seed layer restores (101) at RT and improves critical current density. That experimental dataset is the real contribution and it looks credible—the 80°C switch and the seed-layer effect are exactly what the qualitative story predicts. The writing is clear, the references are appropriate, and the extension to Na and K gives the paper reach.\n\nThe soft spot is the quantitative bridge. The mechanism says load stress changes surface-energy anisotropy enough to flip grain selection, but the DFT supporting data (ref 16) are at ~3% lattice strain. A 5 MPa stack pressure on Li (B≈14 GPa) gives elastic strain ~0.04%, two orders of magnitude smaller. Even allowing for limited creep and plastic deformation, the lattice strain does not approach 3%; the yield strength of Li caps the sustainable stress at ~0.5 MPa unless you invoke strong size effects or stress concentration, and the paper doesn't make that case. Scaling the DFT surface-energy shifts linearly to 0.04% gives a Δγ of ~0.001 J/m², corresponding to a surface-energy density difference of ~10² J/m³ at h=10 μm, while the model's strain-energy term with σy=0.55 MPa and ε_T-ε_c=0.4 gives 10³–10⁴ J/m³. So under realistic strains the phase-field model would predict (101) at RT, contradicting the observed (001) texture. The paper also uses an assumed exponential strain-mobility relation (Eq. S9) with beta set to 1 and no sensitivity analysis. The temperature crossover in Fig. 3A is therefore not secured by the derivation.\n\nThat's a load-bearing gap, not a minor quibble, because the central claim—that load stress causes the texture—depends on it. The experiments are good enough that the empirical correlation remains interesting; the mechanism should be treated as a hypothesis. What would fix it: strain measurements or a defensible argument for local strains, calibrating beta, and a sensitivity analysis.\n\nWho is this for? Battery researchers, especially solid-state and Li-metal community. It deserves a serious referee because the EBSD data and seed-layer result could become a reference point even if the mechanism argument is revised. I would send it to review, with the expectation that the authors address the strain gap. I would not cite it in the next year except perhaps for the EBSD observations.\n\nRecommendation: engage with it, but push on the quantitative link.","headline":"Beautiful EBSD, but the load-stress mechanism is quantitatively unsupported: the DFT surface-energy anisotropy is computed at ~3% strain, while the experiments apply 5 MPa that elastically strains Li by ~0.04%.","tokens_in":16235,"tokens_out":3957,"would_cite":false,"duration_ms":38454,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"In solid-state Li batteries, stacking pressure makes deposited Li adopt a slow-diffusion (001) texture, and a soft amorphous Li-Si seed layer restores the fast (101) texture and improves critical current density.","keywords":["grain selection growth","phase-field model","lithium metal anode","solid-state battery","surface energy anisotropy","critical current density","amorphous Li-Si seed layer","EBSD texture analysis"],"falsifier":"Deposit Li in anode-free solid-state cells at fixed 25 °C while sweeping the stacking pressure from roughly 0.3 MPa to 10 MPa, and map the grain texture by EBSD or XRD: the model predicts a crossover from (101)-dominant to (001)-dominant selection once surface-energy anisotropy crosses a critical threshold. Seeing no pressure-driven texture switch, or a switch at very different pressures or temperatures, would falsify the load-stress anisotropy mechanism.","tokens_in":15085,"feed_emoji":"🔋","tokens_out":11117,"duration_ms":93450,"temperature":0.7,"pith_summary":"This paper proposes a thermodynamic account of which crystallographic texture wins when soft metals like lithium are electrodeposited, and uses that account to explain a known bottleneck of solid-state Li batteries. The central claim is that under the multi-MPa stacking pressure used in solid-state cells, lattice strain raises the surface energy of the fast-diffusing (101) orientation enough that slow-diffusing (001) grains win the growth competition, and this texture is what limits the room-temperature critical current density (the current beyond which the cell fails). The authors support the claim with a phase-field grain-growth model fed by first-principles (DFT) surface energies and diffusion barriers, and with electron backscatter diffraction (EBSD) measurements showing (001)-dominated Li at 25 °C switching to (101)-dominated Li at 80 °C. They then show that an amorphous Li$_x$Si$_{1-x}$ seed layer, which softens the substrate and relieves strain in the Li, restores (101) texture at room temperature and raises the critical current density. If right, the work turns stack pressure into a texture-control parameter and gives a concrete rule for designing seed layers by bulk modulus and amorphicity.","feed_headline":"Soft seed layer restores fast lithium texture in solid-state cells","feed_subtitle":"Stacking pressure favors slow (001) Li grains; an amorphous Li-Si layer flips deposition back to fast (101) texture.","key_machinery":"The load-bearing object is the energy-density difference between two adjacent grains, $\\Delta U_{12} = \\Delta\\Gamma_{\\mathrm{surface}} + \\Delta F_{\\mathrm{strain}}$, with surface term $(E_{S,1}-E_{S,2})/h$ and strain term $\\sigma_y[\\varepsilon_1-\\varepsilon_2]$. The strain term is closed by an exponential relation between intrinsic strain and atomic mobility, $\\varepsilon_{\\mathrm{intrinsic}} = \\varepsilon_c + (\\varepsilon_T-\\varepsilon_c)\\exp(-\\beta D_i/(LR))$, so faster-diffusing grains store less strain energy. This driving force is embedded in an Allen-Cahn-type phase-field model in which grain-boundary mobility $L_q$ follows an Arrhenius law from DFT diffusion barriers and the gradient-energy coefficient $\\kappa_q$ grows with DFT surface energy; layer-by-layer activation of the computational domain mimics progressive electrodeposition and lets the two energy terms compete over deposition time.","core_discovery":"On the paper's own terms, the discovery is that grain selection in electrodeposited soft metals is set by a competition between surface-energy density and mobility-related intrinsic strain energy, and that external load stress tilts the competition by changing surface-energy anisotropy. In solid-state Li cells the winner under load at room temperature is the (001) orientation, which has the highest Li self-diffusion barrier (0.14 eV), so the deposited metal is kinetically constrained; at elevated temperature, strain-energy relief from faster diffusion overtakes surface energy and (101) grains dominate. The paper further claims that a 500 nm amorphous Si layer, lithiating to Li$_x$Si$_{1-x}$ with 0.50 < x < 0.79 and bulk modulus below 30 GPa, softens the interface enough to keep (101) selection at 25 °C, and demonstrates improved rate capability in anode-free full cells.","pith_inferences":["The paper does not say this, but the mechanism implies that stack pressure is a tuneable texture switch: a pressure sweep at fixed temperature should move a cell from (101)-dominant to (001)-dominant deposition, so texture could be engineered by pressure alone.","The same strain-relief logic suggests that any low-modulus, lithiated amorphous interlayer, not only Si, should restore fast-diffusion texture; the paper's four design criteria give a screening rule for testing other seed materials.","A direct extension would be to correlate the (101)-to-(001) texture crossover with critical current density in the same cell, which would test whether texture, rather than interfacial contact alone, is the rate-limiting factor.","Because the crossover temperature depends on the uncalibrated strain-mobility constant, in-situ stress or strain measurements during Li deposition could calibrate the model and sharpen its quantitative phase map."],"forward_implications":["At room temperature, anode-free solid-state Li cells on bare Cu should deposit a (001)-textured layer whose high diffusion barrier limits plating and stripping kinetics, which is why their critical current densities stay below roughly 1.5 mA/cm².","Heating the cell to 80 °C flips the winning texture to (101), and the paper's EBSD measurements confirm that switch.","Because Na and K surface energies respond far less to lattice strain, their room-temperature solid-state deposition should stay (101)-dominated, which makes Na a promising candidate for anode-free solid-state cells.","A seed layer that is amorphous, electronically conductive, lithophilic, and softer than about 30 GPa in bulk modulus should restore (101) texture at room temperature; the paper demonstrates this with amorphous Li$_x$Si$_{1-x}$ (0.50 < x < 0.79)."],"supporting_citations":[{"why":"First-principles data on how lattice strain changes Li surface energies, the input that makes (001) win under load.","marker":"[16]"},{"why":"First-principles self-diffusion barriers for Li, Na, and K, the input that makes (101) the fast-diffusion orientation.","marker":"[17]"},{"why":"Source of the exponential intrinsic-strain/mobility relation used to convert diffusion barriers into strain-energy differences.","marker":"[15]"},{"why":"Observed (101) texture for Li deposited from liquid electrolyte, used to validate the strain-energy-minimizing regime.","marker":"[20]"},{"why":"Independent observation of (101) Li texture in liquid cells, supporting the same prediction.","marker":"[21]"},{"why":"EBSD imaging showing Na grains grow preferentially along (101) in solid-state cells, validating the prediction for Na.","marker":"[22]"},{"why":"First-principles Li-Si alloy study showing amorphous phases soften below 30 GPa, the basis for the seed-layer design.","marker":"[31]"},{"why":"Measured Li yield strength used to set the strain-energy magnitude in the thermodynamic model.","marker":"[13]"}],"fun_headline_variants":["Amorphous LiSi layer flips grain texture for fast Li deposition","Soft seed layer restores fast Li texture in solid-state cells","Strain energy tips Li grain selection to fast growth","Surface and strain energies duel to set Li grain texture","Amorphous seed restores fast lithium texture under load"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The quantitative predictions rest on the assumed exponential relation between atomic mobility and intrinsic strain relief (Eq. S9), with the fitting constant $\\beta$ set to 1 and strain limits taken as 0.2 without calibration against stress or grain-growth measurements; a different form or different constants would move the predicted 25 °C/80 °C texture switch and the pressure-texture phase map.","fun_headline_variants_meta":{"raw":{"variants":["Amorphous LiSi layer flips grain texture for fast Li deposition","Soft seed layer restores fast Li texture in solid-state cells","Strain energy tips Li grain selection to fast growth","Surface and strain energies duel to set Li grain texture","Amorphous seed restores fast lithium texture under load"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001169,"raw_usage":{"total_tokens":4809,"prompt_tokens":895,"completion_tokens":3914,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":511,"completion_tokens_details":{"reasoning_tokens":3833}},"tokens_in":511,"tokens_out":3914,"duration_ms":28283,"temperature":1.0,"reasoning_tokens":3833,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T19:15:17.186846+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Deposit Li in anode-free solid-state cells at fixed 25 °C while sweeping the stacking pressure from roughly 0.3 MPa to 10 MPa, and map the grain texture by EBSD or XRD: the model predicts a crossover from (101)-dominant to (001)-dominant selection once surface-energy anisotropy crosses a critical threshold. Seeing no pressure-driven texture switch, or a switch at very different pressures or temperatures, would falsify the load-stress anisotropy mechanism.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"First-principles data on how lattice strain changes Li surface energies, the input that makes (001) win under load."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"First-principles self-diffusion barriers for Li, Na, and K, the input that makes (101) the fast-diffusion orientation."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Source of the exponential intrinsic-strain/mobility relation used to convert diffusion barriers into strain-energy differences."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Observed (101) texture for Li deposited from liquid electrolyte, used to validate the strain-energy-minimizing regime."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Independent observation of (101) Li texture in liquid cells, supporting the same prediction."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"EBSD imaging showing Na grains grow preferentially along (101) in solid-state cells, validating the prediction for Na."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"First-principles Li-Si alloy study showing amorphous phases soften below 30 GPa, the basis for the seed-layer design."},{"cited_title":"-H., Sanchez, A.J., Poli, A., Arruda, E.M., Thouless, M.D., and Dasgupta, N.P","cited_arxiv_id":null,"evidence_quote":"Measured Li yield strength used to set the strain-energy magnitude in the thermodynamic model."}],"review_version":1}