{"id":"7c226928-cceb-4ad5-9a92-fcd3931495be","arxiv_id":"2506.11177","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"By tracking strains on micropillars with digital image correlation and controlling electron beam damage, the authors report wood cell wall stiffness up to 42 GPa at zero microfibril angle, the closest-to-model direct measurements to date.","lead":"This paper measures the stiffness of the load-bearing layer of wood cell walls using tiny milled pillars and camera-based strain tracking, reporting values up to 42 GPa that are closer to theoretical predictions than earlier tests. It also shows that the electron beam used for imaging can damage the pillars, which may explain why past wood cell wall measurements were scattered and too low.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 42 GPa DIC modulus is a survivor average: the adaptive ROI excludes kink-band damage zones, lowering measured strain and inflating stiffness; a full-pillar ROI re-analysis is needed to settle it.","rationale":"The paper is a careful experimental study with internal consistency: the orientation dependence, the beam-damage parameter study, and the FEM cross-checks all support the qualitative narrative. The main quantitative claim, however, rests entirely on DIC strains, and the DIC pipeline explicitly removes localized damage regions. This is not a disagreement with the consensus; it is an internal consistency issue. The manuscript flags the adaptive ROI in Sec. 3.2, which is to the authors' credit, but flagging a limitation does not remove its effect on the headline number. A full-ROI re-analysis is feasible from the stored image sequences and would settle the magnitude of the bias. Without it, the safest verdict remains conditional: accept the methodology and the orientation trends, but do not treat 42 GPa as a settled record until the survivor-bias check is reported. I agree with the reader's weakest-assumption identification. My pass adds one concrete detail: Table 2 already contains a relevant internal comparison (E_DIC=42 vs E_FEM=37 for the same 2 kV condition) that makes the predicted bias direction and rough magnitude visible, strengthening rather than weakening the need for the check.","tokens_in":15246,"tokens_out":4676,"duration_ms":58113,"concrete_test":"Re-analyze the stored 2 kV, MFA=0° image stack used for Fig. 10 and Table 2 with a fixed ROI spanning the full pillar width and height, without adaptive exclusion, and recompute the cumulative DIC strain exactly as in Sec. 2.3. If the resulting E drops by more than about 10% (for example, from 42 GPa toward or below the 37 GPa FEM-corrected value), the headline claim must be reworded as the stiffness of the undamaged pillar fraction, and the comparison to micromechanical models becomes ambiguous. Also report the excluded pixel fraction as a function of applied strain to quantify when and how much localized damage is screened out.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is the DIC-derived E = 42±3 GPa for MFA=0° under 2 kV imaging. The manuscript itself states in Sec. 3.2 that 'due to the adaptive ROI, zones with localized damage are excluded from the DIC strain calculation.' That is the load-bearing assumption: the strain denominator is computed only over intact regions, while the stress numerator is the total load on the entire pillar. In compression of fibril-aligned wood, kink bands are the main localization mechanism and accommodate a disproportionate share of the imposed displacement. Removing them from the strain average makes the reported modulus a property of the undamaged fraction, not of the whole pillar, and biases E upward. The same logic affects the 'closest match to models' claim: model predictions (55–69 GPa) describe the intact cell-wall material, so comparing them to a survivor-biased measurement is not a clean test. The absence of a DIC value for the 'No Beam' protocol (Table 2) removes the cleanest control: the no-beam FEM-corrected value is 38 GPa, already below the 2 kV DIC 42 GPa. The FEM-based sink-in correction is a useful independent cross-check, but it cannot correct ROI selection, and the 37 GPa FEM value for the same 2 kV pillars leaves open a roughly 12% systematic gap in exactly the direction survivor bias would produce.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript presents micropillar compression (MPC) tests on Norway spruce S2 cell walls with digital image correlation (DIC) strain measurement using platinum microdots, across four microfibril angles (0°, 20°, 70°, 90°) and three SEM imaging protocols (no beam, 2 kV, 5 kV). The central claim is that DIC-based strain mapping yields E = (42±3) GPa for MFA = 0° at 2 kV, which the authors describe as the highest direct stiffness measurement for wood cell walls to date and as the closest experimental match to micromechanical model predictions. A second central claim is that uncontrolled electron-beam exposure in earlier MPC studies is the primary source of their scatter and mechanical underestimation. The paper also compares DIC strains with indentor-displacement strains corrected by the modified Sneddon approach and by orthotropic FEM sink-in models, and reports consistent MFA-dependent trends in modulus, yield stress, and failure modes.","tokens_in":15557,"tokens_out":6658,"duration_ms":83507,"significance":"If the headline result holds, the paper is significant: it would provide a direct, image-based measurement of S2 stiffness far closer to micromechanical model predictions than previous MPC or nanoindentation data, and it identifies electron-beam dose as a controllable source of variability in a protocol that otherwise suffers from large scatter. Strengths include the detailed experimental protocol, the systematic comparison of three imaging conditions, the independence of the DIC strain measurement from the FEM sink-in corrections, and the internally consistent orientation dependence of stiffness, yield stress, and failure mode. The FEM sink-in corrections come from prior published modeling rather than being fitted to the DIC data, so the central comparison is not circular. The main risk is that the DIC strains are averaged over an adaptive region of interest that excludes localized damage, which could bias the headline 42 GPa modulus upward; the paper itself states this exclusion in Sec. 3.2. The absence of DIC data for the no-beam control further weakens the cleanest beam-damage comparison.","major_comments":[{"comment":"Sec. 3.2 states that 'due to the adaptive ROI, zones with localized damage are excluded from the DIC strain calculation.' This creates a mismatch between the stress and strain definitions: the engineering stress is computed from the total measured force over the full pillar cross-section, while the DIC strain is averaged only over intact regions. In fibril-aligned compression, kink bands (Fig. 8) accommodate a disproportionate share of the imposed displacement, so excluding them makes the strain denominator smaller than the whole-pillar average and biases the modulus upward. This is the direction of the gap between E_DIC = 42 GPa and E_FEM = 37 GPa for the same 2 kV pillars in Table 2. Please re-analyze the image sequences with a full-pillar ROI that includes the damaged zones and report both values, or, if the intended quantity is explicitly an intact-material modulus, state that interpretation and restrict the comparison with the 55–69 GPa model predictions accordingly.","section":"Sec. 3.2"},{"comment":"The no-beam control rows in Table 2 have no E_DIC values, so the cleanest comparison for the beam-damage claim is not available: the no-beam value of 38 GPa is obtained only through the FEM sink-in correction, not through direct strain measurement. In addition, the imaging-protocol comparison is inter-pillar rather than same-pillar: different pillars are assigned to the no-beam, 2 kV, and 5 kV arms, so pre-existing pillar-to-pillar variation is not controlled. The conclusion that beam exposure is the primary source of scatter would be stronger if the authors either provide same-pillar evidence, report the individual pillar-level values and exact n for every cell in Table 2, or otherwise explicitly acknowledge that the comparison is between different pillars.","section":"Table 2 / Sec. 2.2"},{"comment":"The 'closest match' language should be calibrated against the actual numbers: E_DIC = 42 GPa is still 24% below the lower cited model prediction of 55 GPa and 39% below the upper prediction of 69 GPa [12,30], while the 90° value of 7±1 GPa matches the 8 GPa model prediction. The claim that the result is closer than previous experiments is supported by the cited literature values, but the adaptive-ROI bias discussed above could act to inflate the measured modulus, which would worsen the agreement with the models if corrected. Please state the residual gap explicitly and avoid presenting 'closest experimental estimate' as equivalent to model validation.","section":"Sec. 3.2 / Conclusions"},{"comment":"The abstract and conclusion identify uncontrolled electron-beam exposure as the primary source of scatter and underestimation in earlier MPC studies, but this is an inference from comparing the present three imaging protocols with literature values. The manuscript does not quantify the beam dose, scan rate, or imaging history of the cited studies, so 'primary source' is stronger than the evidence supports. Please either soften this to 'a major contributor' or provide a quantitative comparison of imaging conditions in the cited works.","section":"Conclusions"}],"minor_comments":[{"comment":"The header contains the typo 'Nanoidentation'; it should read 'Nanoindentation'.","section":"Table 2"},{"comment":"The text says the most probable MFA was 10° and then refers to pillars with MFA 0°, 20°, 70°, and 90°; clarify whether these are local S2 fibril angles relative to the pillar axis or nominal cutting angles, and how the 10° value maps to the '0°' label.","section":"Sec. 2.1"},{"comment":"The modified Sneddon correction in Eq. (1) uses Poisson's ratio ν, but the value of ν used in the analysis is not stated; please specify it.","section":"Sec. 2.3 / Eq. (1)"},{"comment":"The abstract calls the DIC strain measurement 'model-free', but the stress calculation assumes a circular cross-section and does not explicitly correct for barrelling; consider wording such as 'strain measured directly from image correlation' to avoid overstatement.","section":"Sec. 2.3"},{"comment":"The moving regression window of approximately 0.5% used for yield-stress extraction is introduced without a sensitivity study; please report how σyield changes with the window size.","section":"Sec. 2.3"},{"comment":"The figure caption says each curve averages 'at least 4' micropillars; please give the exact number of pillars for each MFA and protocol cell, and indicate which values are standard deviations of the mean.","section":"Fig. 10 / Table 2"}],"recommendation":"major_revision","confidential_remarks":"The stress-test concern about the adaptive ROI exclusion lands on reading of the manuscript: the paper itself states in Sec. 3.2 that damaged zones are excluded from the DIC strain calculation, and the stress is still computed over the full pillar. I did not find circularity in the central DIC-vs-model comparison, since the DIC strains are measured independently of the FEM sink-in corrections. The missing DIC data for the no-beam control and the inter-pillar design of the beam-damage comparison are secondary but real weaknesses. If the authors can provide a full-ROI re-analysis of their existing image sequences, or explicitly re-frame the reported modulus as an intact-material quantity, a revised version could be suitable for publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a genuinely useful paper. It adds DIC strain mapping to wood cell-wall micropillar compression, and the acceleration-voltage study is the clearest demonstration I’ve seen that SEM beam exposure degrades wood pillars and explains much of the scatter in older MPC data. The protocol details are strong: controlled MFA orientations, low-current FIB milling, microdot deposition, and a serious attempt to separate sink-in from material strain via both Sneddon-type and orthotropic FEM corrections.\n\nThe headline result—42 GPa at MFA=0° under 2 kV imaging—is plausible, but it carries a caveat the authors disclose in Sec 3.2: the adaptive ROI excludes zones with localized damage from the DIC strain calculation. In compression along the fibril direction, kink bands are exactly where much of the deformation concentrates. Excluding them means the strain denominator is averaged over intact material only, so the modulus is a survivor average, biased upward relative to the whole pillar. That matters for the “highest direct measurement” claim. It is not fatal, because the FEM-corrected indentor-displacement value for the no-beam pillars is 38 GPa, close to the DIC 42 GPa and well above the historical ~8 GPa. That independent support tells me the conclusion is not purely an artifact of ROI selection. Still, the authors should report a full-pillar ROI analysis or at least quantify the strain fraction excluded, and soften the “direct” language.\n\nOther soft spots are minor. The beam-damage comparison is inter-pillar rather than same-pillar before/after scanning, so individual variability is not fully controlled. The FEM corrections come from the authors’ own prior modeling, but those are published predictions, not fits to this dataset. No raw data or code is provided, which limits independent checking but is common for this kind of study.\n\nBottom line: this deserves a serious peer review. It will be most useful to people working on micromechanical testing of wood and other anisotropic biocomposites. I’d recommend acceptance after moderate revisions, mainly asking for a robustness check on the ROI choice and a clearer statement that the DIC modulus describes the intact fraction of the pillar.","headline":"Careful, useful experimental study that makes a plausible case that beam damage explains past scatter in wood micropillar compression; the 42 GPa DIC number is a survivor average, but FEM-corrected values independently support the main conclusion.","tokens_in":16099,"tokens_out":2408,"would_cite":true,"duration_ms":28709,"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":"Wood cell-wall stiffness reaches 42 GPa when strain is mapped directly on micropillars, pointing to electron-beam damage as the cause of earlier scatter.","keywords":["micropillar compression","digital image correlation","wood cell wall","S2 layer","microfibril angle","electron beam damage","Norway spruce","sink-in correction"],"falsifier":"Run the same MFA=0° compression with DIC strains computed over the entire pillar height including kink bands as well as over the adaptive ROI that excludes them; if the full-field modulus falls substantially below 42 GPa, the headline value is an intact-material survivor average rather than a whole-pillar modulus. A second check is to test pillars with no continuous SEM scanning and derive the modulus from indentor displacement with the orthotropic tracheid sink-in correction; agreement with 42 GPa would confirm that the DIC value is not an artifact of excluding damaged zones.","tokens_in":15057,"feed_emoji":"🌲","tokens_out":8157,"duration_ms":74829,"temperature":0.7,"pith_summary":"This paper tries to show that the scatter and low stiffness values in earlier micropillar compression tests on wood came mostly from the measurement protocol, not from the material. By depositing tiny platinum dots on micropillars cut from the thick, cellulose-reinforced S2 layer of Norway spruce cell walls, and using digital image correlation (DIC) to read strain directly off the pillar surface, the authors report a stiffness of 42 GPa at zero microfibril angle, the highest direct cell-wall stiffness measurement reported so far and the closest to micromechanical model predictions. They also show that continuous electron-beam scanning degrades the pillars, with 5 kV imaging cutting strength by at least half, which they argue explains the scatter and underestimation in earlier studies. This matters because the S2 layer dominates wood's load-bearing response, so reliable values of its stiffness are the foundation for hierarchical models of wood as a structural material.","feed_headline":"Direct strain mapping puts wood cell-wall stiffness at 42 GPa","feed_subtitle":"Electron-beam damage, not the material, explains why earlier micropillar tests came in low.","key_machinery":"The load-bearing mechanism is the measurement chain: FIB-milled micropillars from the S2 layer, a row of platinum microdots deposited on each pillar surface, and digital image correlation (DIC) that tracks the dots in consecutive SEM images to build incremental strain fields. This bypasses the usual need to convert indentor displacement into pillar strain, which requires a sink-in correction. The authors compare the DIC strain against displacement-based strains corrected by the modified Sneddon formula and by FEM models of increasing realism, an orthotropic half-space and a full tracheid assembly, to judge which global correction is trustworthy. A parameter study of SEM acceleration voltage, with no continuous scanning, 2 kV, and 5 kV, identifies electron-beam damage as the factor that moves measured stiffness and strength across the range seen in earlier publications.","core_discovery":"The paper's central claim is that DIC-based strain mapping, not indentor displacement, is the right yardstick for wood micropillar compression. With this method and controlled imaging at 2 kV, the S2 layer of Norway spruce shows a Young's modulus of $E = (42 \\pm 3)$ GPa at a microfibril angle of 0°, compared with model predictions of 55 and 69 GPa; at 90° the measured $E = (7 \\pm 1)$ GPa matches the model value of 8 GPa. The authors attribute the historical scatter in wood micropillar data to electron-beam damage: pillars continuously scanned at 5 kV lost at least 50% of their strength, and their measured modulus fell to values similar to older literature. They further show that standard isotropic sink-in corrections underestimate penetration into the anisotropic cell-wall substrate, while orthotropic FEM corrections with realistic tracheid geometry bring indentor-displacement strains close to the DIC values. The paper concludes that previous micropillar results spanned the full range from minimally degraded to severely degraded material, and that uncontrolled beam exposure is the primary source of the reported variability.","pith_inferences":["Beyond the paper: because the adaptive ROI excludes localized damage zones, the 42 GPa value is best read as the stiffness of the intact S2 material, while the whole-pillar engineering stiffness under large deformation could be lower.","Beyond the paper: if electron-beam damage is the dominant source of scatter, then protocol standardization, especially limiting continuous SEM scanning and using low acceleration voltages, should make future wood micropillar results reproducible across laboratories.","Beyond the paper: the same DIC-on-microdots approach could be extended to other soft anisotropic biological tissues and to moisture- or temperature-controlled in situ testing, provided the imaging dose per test can be kept low enough to preserve the material."],"forward_implications":["Reported S2 stiffness at a microfibril angle of 0° rises to $(42 \\pm 3)$ GPa under 2 kV imaging, the closest experimental match so far to micromechanical model predictions of 55–69 GPa.","As the microfibril angle increases, stiffness and yield stress fall, with the 90° value of $(7 \\pm 1)$ GPa matching the model prediction of 8 GPa.","Pillars continuously scanned at 5 kV show at least 50% strength reduction, so uncontrolled beam exposure, not sample variation, can explain much of the scatter in earlier wood micropillar studies.","Sink-in corrections built on isotropic half-space assumptions underestimate penetration in anisotropic cell-wall material, while orthotropic tracheid-level FEM corrections reproduce DIC-based strains more closely.","Failure mode switches from fibril-aligned kink bands at low microfibril angles to compressive collapse of fibrillar layers at high angles, consistent with ultrastructural alignment.","The protocol combining low-current FIB milling, microdot deposition, DIC strain mapping, and controlled imaging offers a template for probing soft, anisotropic biological composites at the cell-wall scale."],"supporting_citations":[{"why":"Supplies the modified Sneddon sink-in correction used as the global-displacement baseline.","marker":"[34]"},{"why":"Provides earlier wood micropillar compression and yield-stress modelling that the new results are compared against.","marker":"[38]"},{"why":"Gives earlier micropillar compression values that the paper argues were degraded by beam exposure.","marker":"[39]"},{"why":"Supplies the hierarchical micromechanical model predictions of 69 GPa and 8 GPa used for comparison.","marker":"[30]"},{"why":"Supplies the micromechanical cell-wall model prediction of 55 GPa used for comparison.","marker":"[12]"},{"why":"Provides the X-ray scattering method used to determine the local microfibril angle before pillar orientation.","marker":"[42]"},{"why":"Provides the optical-flow algorithm that computes the displacement fields for DIC strain increments.","marker":"[44]"},{"why":"Provides the classical punch-into-elastic-half-space solution underlying the Sneddon correction.","marker":"[47]"}],"fun_headline_variants":["DIC reveals wood cell wall modulus: 42 GPa without beam damage","Beam damage, not wood, caused low micropillar stiffness readings","Wood cell wall hits 42 GPa when imaging dose is controlled","Strain-mapped micropillars show wood's true 42 GPa stiffness","Controlled imaging puts wood cell-wall modulus at 42 GPa"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The DIC strain average excludes zones with localized damage through an adaptive region of interest, so the headline 42 GPa modulus describes only the intact material that survived long enough to be tracked, not the whole pillar including its damaged regions.","fun_headline_variants_meta":{"raw":{"variants":["DIC reveals wood cell wall modulus: 42 GPa without beam damage","Beam damage, not wood, caused low micropillar stiffness readings","Wood cell wall hits 42 GPa when imaging dose is controlled","Strain-mapped micropillars show wood's true 42 GPa stiffness","Controlled imaging puts wood cell-wall modulus at 42 GPa"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000215,"raw_usage":{"total_tokens":1500,"prompt_tokens":1085,"completion_tokens":415,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":701,"completion_tokens_details":{"reasoning_tokens":319}},"tokens_in":701,"tokens_out":415,"duration_ms":5358,"temperature":1.0,"reasoning_tokens":319,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T04:23:19.943581+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the same MFA=0° compression with DIC strains computed over the entire pillar height including kink bands as well as over the adaptive ROI that excludes them; if the full-field modulus falls substantially below 42 GPa, the headline value is an intact-material survivor average rather than a whole-pillar modulus. A second check is to test pillars with no continuous SEM scanning and derive the modulus from indentor displacement with the orthotropic tracheid sink-in correction; agreement with 42 GPa would confirm that the DIC value is not an artifact of excluding damaged zones.","supporting_citations":[{"cited_title":"Zhang, B","cited_arxiv_id":null,"evidence_quote":"Supplies the modified Sneddon sink-in correction used as the global-displacement baseline."},{"cited_title":"Kl ´ ımek, V","cited_arxiv_id":null,"evidence_quote":"Gives earlier micropillar compression values that the paper argues were degraded by beam exposure."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the hierarchical micromechanical model predictions of 69 GPa and 8 GPa used for comparison."},{"cited_title":"Salm´ en, Micromechanical under- standing of the cell-wall structure, C","cited_arxiv_id":null,"evidence_quote":"Supplies the micromechanical cell-wall model prediction of 55 GPa used for comparison."},{"cited_title":"R¨ uggeberg, F","cited_arxiv_id":null,"evidence_quote":"Provides the X-ray scattering method used to determine the local microfibril angle before pillar orientation."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the optical-flow algorithm that computes the displacement fields for DIC strain increments."}],"review_version":1}