{"id":"7623169d-45de-43df-81e6-9ef155a99684","arxiv_id":"2411.16363","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":3,"one_line_summary":"Experiments and simulations show liquid inclusions in elastomers deform non-uniformly, interact strongly as pairs, and can develop elastic creases at their poles under uniaxial tension.","lead":"This study measured how liquid glycerol droplets embedded in soft PDMS rubber deform and interact when the rubber is stretched, using confocal microscopy images and nonlinear simulations. The results show droplets deform unevenly, feel each other's presence strongly, and can develop small reversible creases at their poles.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Finite-thickness free surfaces are untested: for A=264 µm the inclusion surface is ~0.9A from the 1 mm specimen faces while the simulation boundary is 19A away, so the claimed 'highly accurate' infinite-domain agreement for large inclusions could be fortuitous.","rationale":"The paper has real strengths: direct confocal measurements, full-field FE simulations, and consistency with the expected size-independence from small elasto-capillary numbers. My read is that the main logical hinge is the boundary-value problem used for validation. The central claim that the simulations are 'highly accurate' for all tested specimens requires that the simulated domain faithfully represent the experiments. The manuscript itself flags the domain assumption in Section 3 with 'numerical experiments show...' but gives no supporting data, and the chosen 40A cube is much larger than the actual specimen in the thickness direction for the largest inclusions. The finite-thickness free surfaces are the most concrete mismatch: for A=264 µm, the free surface is less than one radius from the inclusion surface. A free surface in an incompressible elastomer permits lateral relaxation that a remote boundary does not, so image corrections are not automatically negligible. If the proposed slab test shows negligible differences, the concern is settled and the conditional verdict can be upgraded; if not, the paper needs either a finite-specimen simulation or a restriction of the claim to inclusions with A much smaller than the specimen thickness. I do not see a more load-bearing problem: mesh dependence of the crease is worth checking, but the crease is observed experimentally and appears elastic, so it is a secondary quantitative issue; the use of a literature surface tension is mitigated by the explicit demonstration that results are insensitive to γ and r_i in the eCa range studied; and the lack of repeated specimens affects confidence intervals but not the structural validity of a single comparison. Thus the infinite-domain/finite-thickness gap is the key conditional element, and the reader's conditional verdict remains appropriate.","tokens_in":16620,"tokens_out":6189,"duration_ms":64513,"concrete_test":"Rerun the A=264 µm isolated-inclusion case with the actual finite specimen geometry: a 30x4x1 mm slab (or the 10 mm gauge section with appropriate end conditions), traction-free surfaces, the inclusion centered at mid-thickness, and the same FE scheme and material parameters. Compare midplane a/A, b/A, and hoop stretch λΘ(Θ) at λ = 1.5, 2.0, and 2.4 against the 40A cube results. Also rerun with the inclusion offset by ±100 µm in the thickness direction to bracket placement uncertainty. If the slab results differ from the cube results by more than about 3% in a/A or b/A, the infinite-domain validation is insufficient; if they differ by less, the concern is resolved and the central claim is supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's Section 3 replaces each specimen by a 40A cube with affine loading (Eqs. 8-11), justified by 'numerical experiments show that the boundaries of the specimens can be considered to be infinitely far away from the inclusions.' That justification checks the computational domain, not the physical specimen. Real specimens are 30 mm x 4 mm x 1 mm (§2.2). For the largest isolated inclusion, A=264 µm, the inclusion surface is only about 236 µm (0.89A) from the traction-free faces of the 1 mm-thick specimen if centered in thickness; the simulation instead places the nearest boundary 19A = 5.0 mm away. With center-to-free-surface distance H about 1.9A, the free-surface image correction to the strain field near an incompressible inclusion is of order (A/H)^3 ≈ 0.15, so finite-thickness relaxation is not obviously negligible. The quantitative validation in Figs. 7-8 and the summary claim that simulations provide 'a highly accurate description of the deformation of the inclusions in all the specimens that were tested' therefore rest on an untested equivalence between two different boundary-value problems. This is the load-bearing gap: if the physical free surfaces relax the local deformation enough to shift a/A or b/A by even a few percent, the A=264 µm agreement is coincidental and the predictive claim for large inclusions is unsupported.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper combines experiments and full-field simulations to study the finite elastic deformation of initially spherical glycerol inclusions in a PDMS elastomer under uniaxial tension. The experimental part uses fluorescent confocal microscopy to measure the deformed midplane shapes of isolated inclusions of three initial radii (A = 35, 72, 264 µm) and of pairs of inclusions at three orientations (0°, 45°, 90°). The simulations employ the framework of Ghosh and Lopez-Pamies (2022) with a two-term I1-based hyperelastic model for the PDMS matrix, incompressible fluid inclusions, and constant interface surface tension, solved on a cubic domain of side 40A with a modified FE scheme. The authors report good agreement for the evolution of the major and minor semi-axes, size independence in the negligible-elastocapillarity regime, a strongly non-uniform local deformation with large compressive hoop stretches near the poles, and the nucleation of reversible creases. They also connect the validated framework to a dilute homogenization result.","tokens_in":16962,"tokens_out":8256,"duration_ms":78420,"significance":"The study is a valuable quantitative test of a nonlinear theoretical framework for liquid inclusions in elastomers. Its strengths include genuine out-of-sample validation: the elastomer constants (Table 1) are fitted only to unfilled-PDMS uniaxial data and are never refitted to inclusion shapes, and the surface tension is explicitly shown to have no effect in this elasto-capillary regime. The direct observation of reversible creases at inclusion poles, with a critical stretch reproduced in simulation at a nearby value, is a novel and falsifiable result. If the finite-thickness concern below is resolved, the paper would provide a solid basis for using such simulations in homogenization models of liquid-filled elastomers.","major_comments":[{"comment":"The replacement of the physical specimen by a cube of side 40A is not validated for the largest inclusions. For A = 264 µm, the cube side is 10.56 mm, while the actual specimen thickness is only 1 mm; if the inclusion is centered through the thickness, its surface is about 236 µm (0.89A) from the traction-free faces, whereas in the simulation the nearest traction-free surfaces are 20A away. The sentence \"numerical experiments show that the boundaries of the specimens can be considered to be infinitely far away from the inclusions\" checks the computational domain size, not the actual specimen geometry. For an incompressible inclusion near a free surface, image corrections scale roughly as (A/H)^3, and with H/A ≈ 1.9 that correction is of order 0.15, so the effect is not obviously negligible. I request either a simulation of the actual 30 mm × 4 mm × 1 mm slab with traction-free faces for A = 264 µm, or a report of the through-thickness inclusion position together with a convergence study demonstrating that free-surface effects are below the experimental resolution. Without this, the \"highly accurate\" claim in Section 5 is not fully supported.","section":"Section 3, Eqs. (8)-(11)"},{"comment":"The crease nucleation in the simulation is attributed to \"errors inherent to the use of a FE discretization\" acting as imperfections. Because the trigger is numerical noise, the predicted critical stretch (λ ≈ 1.5) and critical hoop stretch (λΘ = 0.71) are potentially mesh-dependent. The comparison with the experimental value (λ ≈ 1.6) and with the half-space threshold (λc = 0.65) is therefore not a quantitative validation unless mesh convergence is established. Please add a mesh-refinement study (e.g., two or more systematically refined discretizations) and show that the crease nucleation stretch and location do not change appreciably.","section":"Section 4.4, Fig. 11"},{"comment":"The central validation claim is stated as \"highly accurate description of the deformation of the inclusions in all the specimens that were tested,\" but the comparison in Figs. 8 and 13 is only visual and no error bars or quantitative error measures are reported. Because the abstract and conclusions make a strong quantitative claim, please report the number of repeated tests and the maximum/average relative error in a/A and b/A, or otherwise provide a statistical measure of agreement.","section":"Section 4.3, Figs. 8 and 13"}],"minor_comments":[{"comment":"The figure labels and captions use \"mm\" where \"µm\" is intended (e.g., \"A = 72 mm\", \"A = 87 mm\", \"D = 62 mm\"). Please correct these units and verify the scale-bar lengths for each panel.","section":"Fig. 7 and Fig. 12"},{"comment":"The hoop stretch λΘ is defined only in the midplane e1-e2 plane. Since the inclusion is three-dimensional, there is also an out-of-plane hoop direction; please clarify that the reported λΘ is the midplane hoop stretch and explain why this is the relevant measure for the observed creases.","section":"Eq. (16) and Section 4.3"},{"comment":"The crease observation is reported for a single specimen (A = 72 µm). Please state how many specimens were examined for creasing and whether the critical stretch was reproducible, or temper the claim accordingly.","section":"Section 4.4"},{"comment":"In the reference for Poulain et al. (2017), the author list reads \"Lopez-Pamies, Ravi-Chandar\" without given initials; it should be \"O. Lopez-Pamies, K. Ravi-Chandar\" or similar.","section":"References"},{"comment":"The residual stress ri is estimated from Laplace's law as ri = -2γ/A, but the sign convention for pressure in the fluid is not defined before Eq. (14). A brief statement of the sign convention would improve clarity.","section":"Section 4.2"}],"recommendation":"major_revision","confidential_remarks":"The paper is a strong fit for the journal and the experimental/simulation dataset is valuable. The main risk is the finite-thickness approximation for the largest inclusions, which is fixable with additional simulations or through-thickness position data. The crease mesh-sensitivity issue should also be addressed. I see no concerns about novelty or attribution."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Kamal, here's my read on 2411.16363.\n\nThe paper is a solid next installment in the Lopez-Pamies program: confocal in-situ measurements of glycerol inclusions in PDMS under uniaxial tension, with full-field FE simulations using their 2022 framework. The genuinely new content is local: hoop stretches along the inclusion interface, deviation from ellipsoidal shape via the Delta-lambda_theta measure, orientation-dependent pair interactions at 0, 45, and 90 degrees, and a reversible crease at the inclusion pole. The experimental discipline earns credit: elastomer constants are fitted to unfilled-matrix data and never refitted to inclusion shapes, so the inclusion comparisons are out-of-sample predictions. The surface-tension effect is checked directly and shown to be negligible in this regime. Heavy self-citation is fine here; the framework is theirs and the test is genuine.\n\nI agree with the reader's conditional verdict on the qualitative claims. The quantitative claim — \"highly accurate description of the deformation in all the specimens\" — is where I share the stress-test's concern. The simulations replace the 30x4x1 mm specimen with a 40A cube. For A=264 um, the inclusion surface sits about 0.9A from the traction-free faces of the 1 mm thickness, while the simulation puts the nearest boundary at 19A. The paper's justification tests the computational domain, not the physical geometry. The image correction from a free surface at H ~ 1.9A scales as (A/H)^3 ~ 0.15, which is not obviously negligible. If the A=264 um agreement in Fig. 8 is real, that is reassuring, but without a simulation of the actual slab geometry — or at least a demonstration that hoop stretch and semiaxes are insensitive to a free surface at that distance — the 'highly accurate' claim is only partially supported. This is fixable and should be requested.\n\nOther soft spots are minor: single specimens per condition without error bars, a literature value for surface tension (though it demonstrably does not matter here), and a simulated crease onset at lambda=1.5 versus 1.6 presented without a mesh-convergence study. The explanation that FE discretization errors act as imperfections is a bit hand-wavy.\n\nWho benefits: anyone working on soft composites, liquid-filled elastomers, or crease nucleation in soft materials. The technique and pair-interaction data are worth having. I would cite it if I worked in this area.\n\nRecommendation: send to peer review. It deserves referee time. The requests should be specimen repeats with error bars, a finite-thickness check for large inclusions, and mesh-convergence for the crease. Achievable and they would firm up the central claim.","headline":"A genuinely out-of-sample validation of the Lopez-Pamies liquid-inclusion framework, with new local-deformation data and a reversible pole crease; the 'highly accurate' claim is qualified by an unaddressed finite-thickness effect for the largest inclusions.","tokens_in":17494,"tokens_out":3946,"would_cite":true,"duration_ms":38295,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["74B20","74Q20","74S05"],"pacs":[],"model":"deepseek-v4-flash","headline":"Experiments and simulations together show that liquid inclusions in a stretched elastomer deform highly non-uniformly and can develop creases at their poles.","keywords":["liquid inclusions","elastomers","finite deformation","elasto-capillarity","surface tension","creasing","confocal microscopy","homogenization"],"falsifier":"Perform uniaxial tension tests on specimens with identical inclusion radius and spacing but different thicknesses $B$, down to a few times $A$; if the measured inclusion axes, hoop stretch distribution, or crease threshold change measurably with $B$, the infinite-domain assumption that carries the simulations would be falsified.","tokens_in":16398,"feed_emoji":"💧","tokens_out":5365,"duration_ms":49659,"temperature":0.7,"pith_summary":"The paper tries to establish that full-field simulations from a recently developed theoretical framework can describe the nonlinear elastic deformation of liquid inclusions embedded in an elastomer, quantitatively and locally. It tests this on PDMS specimens containing isolated glycerol droplets and pairs of droplets in three orientations, all under uniaxial tension in the regime where elasto-capillarity is negligible. Confocal fluorescent microscopy provides direct measurement of the inclusion shape, and the simulations match the measured evolution of the inclusion axes and shapes. The central payoff would be a validated predictive tool for the local mechanics of liquid-filled elastomers, including the non-uniform deformation and the nucleation of creases at inclusion poles.","feed_headline":"Liquid droplets in elastomers deform unevenly and crease at the poles","feed_subtitle":"Full-field simulations reproduce confocal measurements of isolated and paired glycerol inclusions under stretch.","key_machinery":"The argument rests on a coupled Lagrangian equilibrium framework in which the elastomer is an incompressible non-Gaussian hyperelastic solid, the liquid inclusion is an incompressible elastic fluid with residual hydrostatic stress, and the interface is a hyperelastic surface with constant surface tension. For each experiment the domain is taken as a cube of side $40A$ with affine displacement boundary conditions imposing the measured macroscopic stretch $\\lambda$, and the equilibrium PDEs are solved with a finite-element scheme. The local response is characterized through the hoop stretch $\\lambda_\\Theta$ along the interface and the deviation measure $\\Delta\\lambda_\\Theta$ that compares the actual hoop stretch with that of a uniformly deforming ellipsoid; the crease is detected by the loss of symmetry at the pole in the finite-element solution.","core_discovery":"The paper claims that the simulations provide a highly accurate description of the deformation of the inclusions in all specimens tested, for isolated inclusions and for pairs oriented at 0, 45, and 90 degrees to the load. The deformation of liquid inclusions is significantly non-uniform, with hoop stretches tensile at the equator and compressive near the poles; the compressive hoop stretch can become large enough to nucleate a crease at the pole, which the simulation reproduces at a macroscopic stretch close to the experimentally observed value. The deformation is independent of inclusion size because the elasto-capillary numbers are small, and the presence of a neighboring inclusion modifies the deformation strongly and in an orientation-dependent way, including mutual shielding for pairs aligned with the load.","pith_inferences":["If the critical condition is indeed the local compressive hoop stretch at the pole, the same framework should predict crease onset for other inclusion shapes, spacings, and multiaxial loads; the paper does not test this generality.","The observed elastic crease that disappears on unloading suggests a reversible surface-pattern mechanism, potentially useful for switchable optical or wetting properties of liquid-filled elastomers; this is an extension beyond the paper.","For specimens whose thickness is comparable to the inclusion size, the infinite-domain approximation that carries the simulations will break down, and a finite-thickness simulation with traction-free faces would be a natural test of how much the local fields change.","Extending the same experimental protocol to the elasto-capillarity-dominated regime of smaller inclusions should reveal whether the crease threshold shifts with surface tension; the paper leaves this as an obvious next step."],"forward_implications":["The verified framework can be used to generate full-field local deformation data for a wide range of inclusion arrangements and loading histories without new experiments.","The simulation results can feed homogenization-based effective stored-energy functions for suspensions of liquid inclusions, including the dilute-limit correction function $H(\\mathbf{F})$ presented in the paper.","Crease nucleation at inclusion poles gives a local criterion for failure or patterning in liquid-filled elastomers: the critical macroscopic stretch and the associated critical hoop stretch identify a threshold that can be compared across geometries.","Orientation-dependent interaction between inclusions means that local strain concentrations can be designed by arranging inclusions in specific patterns relative to the loading direction."],"supporting_citations":[{"why":"It supplies the theoretical framework and finite-element implementation used for all the full-field simulations.","marker":"Ghosh and Lopez-Pamies (2022)"},{"why":"It provides the non-Gaussian I1-based hyperelastic stored-energy function used to model the PDMS elastomer.","marker":"Lopez-Pamies (2010)"},{"why":"It provides prior evidence that this stored-energy form accurately describes PDMS Sylgard 184, supporting the fitted material constants.","marker":"Poulain et al. (2017)"},{"why":"It is the source of the estimated surface tension gamma = 0.014 N/m at the elastomer/inclusion interface.","marker":"Style et al. (2015a)"},{"why":"It establishes the small-deformation limit and the elasto-capillary negligibility criterion used to justify ignoring surface-tension effects at the tested inclusion sizes.","marker":"Ghosh et al. (2023b)"},{"why":"It provides the dilute-limit analytical homogenization result that the simulations are compared with in the construction of effective stored-energy functions.","marker":"Lefèvre and Lopez-Pamies (2017a,b)"}],"fun_headline_variants":["Liquid inclusions crease at poles under stretch","Simulations match experiments on liquid-filled elastomers","Neighboring droplets change deformation in soft solids","Droplet creasing at poles reproduced by simulations"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the boundaries of the actual thin specimens are effectively infinitely far from the inclusions, so a cube of side $40A$ with affine boundary conditions reproduces the experiment; for the largest inclusions the simulation cube is much thicker than the real 1 mm specimen, so finite-thickness effects are assumed negligible.","fun_headline_variants_meta":{"raw":{"variants":["Liquid inclusions crease at poles under stretch","Simulations match experiments on liquid-filled elastomers","Neighboring droplets change deformation in soft solids","Droplet creasing at poles reproduced by simulations"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000946,"raw_usage":{"total_tokens":4064,"prompt_tokens":997,"completion_tokens":3067,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":613,"completion_tokens_details":{"reasoning_tokens":3007}},"tokens_in":613,"tokens_out":3067,"duration_ms":22085,"temperature":1.0,"reasoning_tokens":3007,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T13:13:17.404600+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Perform uniaxial tension tests on specimens with identical inclusion radius and spacing but different thicknesses $B$, down to a few times $A$; if the measured inclusion axes, hoop stretch distribution, or crease threshold change measurably with $B$, the infinite-domain assumption that carries the simulations would be falsified.","supporting_citations":[{"cited_title":"Elastomers filled with liquid inclusions: Theory, numerical implemen- tation, and some basic results","cited_arxiv_id":null,"evidence_quote":"It supplies the theoretical framework and finite-element implementation used for all the full-field simulations."},{"cited_title":"A new I1-based hyperelastic model for rubber elastic materials","cited_arxiv_id":null,"evidence_quote":"It provides the non-Gaussian I1-based hyperelastic stored-energy function used to model the PDMS elastomer."},{"cited_title":"Damage in elastomers: Nucleation and growth of cavities, micro-cracks, and macro-cracks","cited_arxiv_id":null,"evidence_quote":"It provides prior evidence that this stored-energy form accurately describes PDMS Sylgard 184, supporting the fitted material constants."}],"review_version":1}