{"id":"3075b1b9-380e-44d6-a1a8-4697ebe05d39","arxiv_id":"2412.05910","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"Swollen elastomers microphase separate into finite domains whose size scales as Y^{-1/2} and whose transition temperature decreases linearly with Young's modulus Y, according to a phase-field model with nonlocal elasticity.","lead":"A nonlocal coarse-grained elasticity model predicts that solvent-swollen elastomers form stable micro-sized domains, with domain size shrinking as the inverse square root of stiffness and the microphase-separation temperature falling linearly with stiffness. The predictions match recent experiments on PDMS elastomers and offer a design rule for fabricating patterned soft materials.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Central scalings rest on the unverified h = nξ assumption with n fitted to the same Λ data; absolute domain size and Tmicro slope are not parameter-free predictions.","rationale":"The reader's weakest-assumption analysis correctly identifies h = nξ as load-bearing. I agree that the scalings Λ~Y^{-1/2} and Tmicro∝Y hinge on this assumption, and that n is fitted to the same data, so the absolute domain size is not independently predicted. However, I find an additional soft spot in the Tmicro comparison: because ψ0 is taken from experimental swelling data and the parameters a, b, T* are fitted, the linearity in Fig. 2(b) is not a free prediction of the nonlocal-elasticity mechanism alone. This strengthens the case for a conditional verdict, but it does not overturn the paper's central physical picture: the Y^{-1/2} slope in Fig. 2(a) is a genuine, non-trivial success that would fail if h were Y-independent, and the kernel-independence proof in SM Sec. V adds real support. I see no internal inconsistency or mathematical error in the derivation of Eq. (8) or Eq. (12). Therefore I do not recommend changing the reader's verdict; the concern is real but is exactly the kind of assumption that warrants 'conditional' rather than 'accept'. My partial disagreement with the reader is that the Tmicro comparison carries the additional fitted-input issue, making the quantitative claim even more conditional than the h~ξ issue alone would suggest.","tokens_in":24216,"tokens_out":23784,"duration_ms":232016,"concrete_test":"Fit the static structure factor S(q) from the experiments of Ref. [24] (or from new small-angle scattering measurements on the same PDMS/solvent system) to Eq. (S38), letting h, M, and κ be free parameters for each stiffness Y. Then plot the best-fit h(Y) against the independently estimated network mesh size ξ(Y) from Eq. (S21) (or from rheological/swelling measurements). If h(Y)/ξ(Y) is not a single constant n across all Y (within experimental error), the h~ξ assumption is falsified. Alternatively, if scattering data are unavailable, use the published Λ(Y) to determine q_m=2π/Λ, fit n to the softest and stiffest samples, and then check whether the same n predicts the intermediate Λ and the Tmicro(Y) slope from Eq. (10) without re-fitting T* and a; failure of this cross-validation would show the quantitative agreement is overfit.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's two headline scalings—Λ~Y^{-1/2} (Eq. 12) and Tmicro linear in Y (Eq. 10)—both follow from the single assumption h = nξ with ξ~(B/Y)^{1/2} (Eq. 11) and M~Y. This assumption is motivated only by the dimensional statement that the elastic continuum holds at scales 'much larger than ξ, but proportional to it'; no derivation is given. The proportionality constant n=110 is fitted to the very domain-size data the theory is supposed to explain, so the absolute magnitude of Λ is not an independent prediction. If h were independent of Y (as in the 1D nonlocal model of Ref. [31]), then q_m^{-2} ~ h^2/ln(Mh^2/κ) would vary only logarithmically with Y, contradicting the observed decade-scale variation in Λ; thus the Y^{-1/2} slope is a genuine conditional success. A second soft spot is the Tmicro comparison: Eq. (10) contains 3bψ0^2, and ψ0=φ0-φ* is taken from experimental swelling data at 60°C; with b=2 kPa and φ0 varying from ~0.26 to ~0.72, this term alone shifts Tmicro by tens of K, so the apparent linearity in Fig. 2(b) depends on the empirical φ0(Y) relation and the fitted T*, a, b. The nonlocal mechanism itself is not invalidated, but the quantitative agreement is softer than a parameter-free test.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript proposes a phase-field model for microphase separation in solvent-swollen elastomers, in which the elastic response is nonlocal: deformations occurring on scales below a coarse-graining length h do not stress the elastomer. The effective Fourier-space interaction is F_q = a(T-T*) + κ q^2 + M exp(-h^2 q^2), which acquires a minimum at finite q when γ = M h^2/κ > 1, giving stable finite-sized domains. Identifying h with nξ, where ξ is the network strand end-to-end distance and ξ ~ (3B/Y)^{1/2} from rubber elasticity, the paper predicts Λ ~ Y^{-1/2} and a linear decrease of the microphase-separation temperature Tmicro with Young's modulus Y. These predictions are compared with the experiments of Fernández-Rico et al. (Nat. Mater. 2024). The paper also constructs mean-field phase diagrams with uniform, droplet, and stripe phases, supported by numerical energy minimization.","tokens_in":24509,"tokens_out":8392,"duration_ms":83802,"significance":"The model provides an elegant and analytically transparent mechanism for elasticity-controlled microphase separation, with the nonlocal elastic energy giving a physical justification for the emergent long-range interaction. The scaling laws are concrete and falsifiable, and the kernel-independence proof in the Supplemental Material is a notable strength. The numerical implementation is openly available, and the connection to Landau–Brazovskii theory and block-copolymer phase diagrams places the work in a broad context. However, the quantitative agreement with experiments is heavily dependent on fitted parameters (n, T*, φ*, a, b), and the phase diagram does not reproduce the bicontinuous morphologies observed in stiff elastomers. The core scaling exponent Λ ~ Y^{-1/2} is a genuine conditional prediction, but the magnitude of Λ is set by the fitted n, and the Tmicro comparison is largely a consistency check.","major_comments":[{"comment":"The absolute magnitude of the predicted domain size Λ is fixed by the fitting parameter n=110 in h=nξ, and n is inferred from the same domain-size data that the theory claims to explain. Consequently, the good agreement for the absolute values of Λ in Fig. 2(a) is partly by construction. The Y^{-1/2} slope is a genuine prediction of the h∝ξ assumption, but the manuscript should state explicitly that only the slope is predicted, not the overall magnitude. Please provide an independent estimate or constraint on n (e.g., from the number of strands per domain) or report the uncertainty in n obtained from the fit.","section":"Comparison to experiments, Eq. (12) and Fig. 2(a)"},{"comment":"Equation (10) is evaluated using T*=70°C, φ*=0.2, a=0.025 kPa/K, and b=2 kPa, with T* and φ* estimated from the Tmicro data and a and b adjusted to match the experiments. Since the crosses in Fig. 2(b) also use the experimental φ0(Y) relation, the apparent linear decrease of Tmicro with Y is a consistency check rather than a parameter-free prediction. Please quantify the sensitivity of the crosses to the fitted parameters and, if possible, determine T*, φ*, a, and b from independent swelling or phase-equilibrium data.","section":"Comparison to experiments, Eq. (10) and Fig. 2(b)"},{"comment":"The model predicts uniform, droplet, and stripe phases, but the experimental system exhibits bicontinuous microstructures in the stiffest samples. The authors acknowledge this and attribute it to shear or nonlinear effects, which is a reasonable limitation. However, the abstract claims that the analytical phase diagrams \"capture the richness of the microphase morphologies,\" which is stronger than what the model delivers. The text should either temper that claim or provide a concrete mechanism (e.g., the fluctuation-induced gyroid scenario mentioned in SM Sec. IV) with a qualitative parameter estimate for the occurrence of bicontinuous structures.","section":"Phase diagram, Fig. 3(a) and surrounding text"}],"minor_comments":[{"comment":"The phrase \"we chooseh as a multiple\" should read \"we choose h as a multiple\" (missing space).","section":"Fig. 1 caption"},{"comment":"The entry for the quartic coefficient b reads \"2 kPa K^{-1}\"; the correct units are kPa, since b appears in the free-energy density as (1/4)bψ^4 with ψ dimensionless.","section":"Table I"},{"comment":"The statement \"Tmicro decreases linearly with the modulus M\" is strictly true only when ψ0 is held fixed; in the experimental comparison ψ0 varies with Y through φ0. Please rephrase as \"for fixed ψ0\" or clarify the dependence.","section":"After Eq. (10)"},{"comment":"The notation \"B µE\" contains an awkward space; it should be written as a single symbol, e.g., \"BμE\".","section":"Supplemental Material, Sec. IV"},{"comment":"The sentence \"With the choice h=nξ, the parameter γ is independent of Y\" would benefit from a brief explanation, since γ = M h^2/κ involves both M~Y and h^2~Y^{-1}.","section":"Comparison to experiments, paragraph before Eq. (12)"}],"recommendation":"major_revision","confidential_remarks":"The paper is a solid phenomenological contribution within the scope of the journal, and the core mechanism is plausible. The main reservations are the parametric nature of the agreement with experiments and the incomplete morphological comparison. I would encourage the editors to ask for a revised version that clearly separates fitted parameters from predicted quantities, quantifies the sensitivity of the results, and tempers the morphological claims. The technical derivation itself is sound and the paper is likely to be of interest to the soft-matter community."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know two things about this one. First, it does something genuinely useful: it takes the nonlocal-elasticity idea from Qiang et al. and turns it into a 3D weak-segregation theory with two sharp scaling predictions, Λ ~ Y^{-1/2} and Tmicro linear in Y, and proves in the SM that those scalings hold for any reasonable coarse-graining kernel. Second, the two predictions are conditional on the assumption that the coarse-graining length h is proportional to the network strand spacing ξ ~ Y^{-1/2}. The proportionality constant n is fitted to the same domain-size data, so the absolute magnitude of Λ is not a parameter-free prediction.\n\nWhat I like: the derivation is clear, the SM is unusually complete, and the numerical minimization supports the one-mode phase diagram. The authors are honest about what they have not captured, and they give proper credit to prior work, including the 1D strong-segregation model and de Gennes's scaling. The code is public, which lets a reader check the numerics.\n\nWhere it gets soft: the h = nξ relation is asserted on dimensional grounds, not derived, and n=110 is fit to Fig. 2(a). The Tmicro comparison is even softer, because the 3bψ0^2 term in Eq. (10) is significant, and ψ0 comes from experimental swelling data plus fitted T*, a, and b. So the apparent linearity in Fig. 2(b) is partly built in. The phase diagram also fails to give the bicontinuous structures seen in stiff samples; the authors suggest shear or nonlinearity, but that's a gap. And the experimental points carry no error bars, which makes the 'good agreement' hard to judge quantitatively.\n\nNone of this kills the central mechanism. The scaling laws are genuine conditional predictions, and the model is a real advance over the 1D treatment. I would send this to a serious referee. For revision, I'd want the authors to discuss the h assumption more critically, run a sensitivity analysis on n, and say more about the bicontinuous phase. Anyone working on elastic phase separation, patterned elastomers, or nonlocal elasticity will get value from it.","headline":"A clean scaling theory for elastic microphase separation; the central scalings are conditional on a fitted h, but the mechanism holds up.","tokens_in":25074,"tokens_out":4193,"would_cite":true,"duration_ms":42059,"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":"A swollen elastomer forms microdomains when molecular demixing and mesoscopic elasticity act at incompatible length scales; predicted size scales as Y^(−1/2) and ordering temperature drops linearly with stiffness, matching PDMS data.","keywords":["microphase separation","elastomer","nonlocal elasticity","phase-field model","polymer network","Landau-Brazovskii free energy","domain-size scaling","rubber elasticity"],"falsifier":"On the same PDMS samples, measure the structure-factor peak position $q_m$ by small-angle scattering during a temperature quench and independently determine the strand length $\\xi$ from swelling or rheology. The model requires the product $q_m\\xi$ to be a stiffness-independent constant equal to $\\sqrt{\\ln\\gamma}/n$ with $n = 110$, so $q_m$ must grow as $Y^{1/2}$, and requires $\\Lambda$ to keep following $Y^{-1/2}$ beyond the fitted 10–800 kPa range. A drift of $q_m\\xi$ with stiffness, or a saturation of $\\Lambda$ at extreme stiffnesses, would falsify the $h = n\\xi$ premise and the scaling laws built on it.","tokens_in":23967,"feed_emoji":"〰️","tokens_out":24980,"duration_ms":205150,"temperature":0.7,"pith_summary":"This paper claims that microphase separation in an elastomer swollen with a solvent can arise purely from the disparity between two length scales: polymer-solvent demixing acts at molecular distances, while the polymer network responds elastically only to deformations larger than a mesoscopic coarse-graining length $h$. The model encodes this in a nonlocal elastic energy that, coupled to the composition field, produces the effective interaction $F_q = a(T-T^*) + \\kappa q^2 + M e^{-h^2 q^2}$ in Fourier space; when the dimensionless elastocapillary parameter $\\gamma = Mh^2/\\kappa$ exceeds 1, this interaction is minimized at a nonzero wavenumber, so the system forms stable finite-sized domains instead of fully separating. Using rubber elasticity to tie $h$ to the network strand end-to-end distance $\\xi \\sim Y^{-1/2}$, the paper predicts the domain size scales as $\\Lambda \\sim Y^{-1/2}$ and the microphase separation temperature decreases linearly with the dry Young's modulus $Y$, both in good agreement with the PDMS experiments it sets out to explain. If correct, the model makes elastic stiffness the practical knob for stable, patterned elastomers, and places the phenomenon in the same universality class as block-copolymer microphase separation.","feed_headline":"Stiffness sets the pattern size in demixing elastomers","feed_subtitle":"Domain size shrinks as Y^(−1/2) while ordering temperature falls linearly with stiffness; PDMS experiments agree.","key_machinery":"The central object is the nonlocal elastic energy density $w(\\varepsilon) = \\frac{\\lambda}{2}(\\operatorname{tr}\\varepsilon)(\\operatorname{tr}\\bar\\varepsilon) + \\mu\\operatorname{tr}(\\varepsilon\\bar\\varepsilon)$, in which the stress-producing strain $\\bar\\varepsilon$ is the convolution of the local strain $\\varepsilon$ with an isotropic, normalized coarse-graining kernel $K_h$ of width $h$ (a Gaussian in the main text); only deformations larger than $h$ stress the network. Coupled to the order parameter $\\psi$ through material conservation, $\\operatorname{tr}\\varepsilon \\approx -\\phi_*^{-1}\\psi$, this energy becomes the Fourier-space binary interaction $F_q = a(T-T^*) + \\kappa q^2 + M e^{-h^2 q^2}$, with $M = (\\lambda+2\\mu)/\\phi_*^2$ the rescaled longitudinal modulus of the swollen elastomer. The dimensionless ratio $\\gamma = Mh^2/\\kappa$ — an inverse elastocapillary number — decides whether a microphase forms: $\\gamma > 1$ puts the minimum of $F_q$ at $q_m = h^{-1}\\sqrt{\\ln\\gamma}$, giving domains of size $\\Lambda \\sim 2\\pi/q_m$, while $\\gamma = 1$ is the Lifshitz point. The stiffness scalings then come from rubber elasticity, $h = n\\xi$ with $\\xi \\sim (3B/Y)^{1/2}$ and $M \\sim \\frac{1}{3}\\phi_*^{-5/3}Y$, which makes $\\gamma$ independent of $Y$ and yields $\\Lambda \\sim Y^{-1/2}$ with $T_{\\rm micro}$ linear in $Y$.","core_discovery":"On its own terms, the paper establishes that the length-scale gap between thermodynamics and elasticity is sufficient, by itself, to select a finite pattern size in a swollen elastomer. The constitutive relation is nonlocal: the strain field is coarse-grained with a normalized kernel of width $h$, so deformations below $h$ do not stress the network, and this elastic energy is coupled to the local polymer volume fraction through material conservation, $\\operatorname{tr}\\varepsilon \\approx -\\phi_*^{-1}\\psi$. The resulting quadratic free energy in Fourier space, $F_q = a(T-T^*) + \\kappa q^2 + M e^{-h^2 q^2}$, pits an interfacial term favoring long wavelengths against an elastic term favoring short wavelengths; for $\\gamma = Mh^2/\\kappa > 1$ the competition yields a minimum at $q_m = h^{-1}\\sqrt{\\ln\\gamma}$, selecting domains of size $\\Lambda \\sim 2\\pi/q_m$. Setting the coarse-graining length by rubber elasticity, $h = n\\xi$ with $\\xi \\sim (3B/Y)^{1/2}$, gives $\\Lambda \\sim Y^{-1/2}$ (Eq. 12), and linear-stability analysis gives $T_{\\rm micro} = T^* - a^{-1}[3b\\psi_0^2 + M\\gamma^{-1}(1+\\ln\\gamma)]$ (Eq. 10), a linear decrease of the ordering temperature with stiffness. These scalings, together with the predicted phase diagram of uniform, stripe, droplet, and inverted-droplet phases (of Landau-Brazovskii type), are the claims the paper puts against experiment.","pith_inferences":["The constant n = 110 is fitted to the same domain-size data, so the model's absolute Λ is not predicted from independent inputs; measuring the coarse-graining scale directly — say from the structure-factor peak width or the correlation length — would reveal whether n is a universal material constant or an absorbent fitting parameter.","Because this model predicts S(0) ≠ 0 where an earlier theory of crosslinked polymer blends predicted S(0) = 0, small-angle scattering on such blends would be a direct experimental discriminator of whether their elastic response is genuinely nonlocal.","The paper's mean-field phase diagram does not yield the bicontinuous structures seen at high stiffness, and the authors attribute this to neglected shear or nonlinear effects; a testable extension is to impose a controlled pre-strain on a stiff sample and look for the droplet phase converting into bicontinuous morphology.","The free energy reduces to Landau-Brazovskii form, so thermal fluctuations of the kind that suppress the mean-field critical point in other modulated-phase systems should act here as well; a numerical or experimental check would be whether the transition at T'_* becomes weakly first-order and whether gyroid-like bicontinuous phases appear at high stiffness."],"forward_implications":["Pattern size becomes tunable by stiffness alone: Λ ~ Y^(−1/2) across the 10–800 kPa range of the PDMS experiments, so more densely crosslinked networks produce proportionally finer microdomains.","The microphase separation temperature falls linearly with Young's modulus, so stiffer elastomers need deeper temperature quenches before ordering begins — a quantitative constraint for fabricating patterned elastomers.","Because γ = Mh²/κ > 1 marks the Lifshitz point, the model predicts first-order transitions from the uniform state into droplet or stripe phases, with a phase diagram of the same generic topology as block-copolymer melts.","Written per network strand, the predictions are Λ ~ N^(1/2) and a T_micro linear in N^(−1) — the same scalings predicted earlier for crosslinked polymer blends, which the paper reads as evidence that such blends may also respond nonlocally to stress.","The static structure factor S(q) ~ 1/F_q peaks at the selected wavenumber and grows smoothly as the temperature approaches T_micro, so a quench should show scattering intensity rising at a fixed wavenumber, as the experiments show."],"supporting_citations":[{"why":"Supplies the experimental PDMS data — domain sizes, microphase temperatures, and morphologies — that the model is fitted against and must reproduce.","marker":"[24]"},{"why":"Introduces the 1D nonlocal-elasticity model of the same experiments that motivates this work and with which its scaling laws and transition order are contrasted.","marker":"[31]"},{"why":"Provides the nonlocal-elasticity framework from which the coarse-grained stress-strain relation, Eqs. (1)–(3), is taken.","marker":"[32]"},{"why":"Gives the theory of phase transitions in polymer gels, including the relation M ~ (1/3)φ_*^(−5/3)Y between swollen longitudinal modulus and dry Young's modulus used in the experimental comparison.","marker":"[28]"},{"why":"Defines the rescaled longitudinal modulus M of polymer gels that enters the elastic contribution to the free energy.","marker":"[74]"},{"why":"Establishes the scaling of the strand end-to-end distance ξ with the dry Young's modulus Y, the source of the h ~ Y^(−1/2) dependence behind Λ ~ Y^(−1/2).","marker":"[84]"},{"why":"Supplies the estimate of the interfacial parameter κ ~ k_BT/ℓ used to set γ and the absolute domain size.","marker":"[87]"},{"why":"Provides the block-copolymer microphase separation theory whose phase-diagram topology the model is shown to reproduce.","marker":"[92]"},{"why":"Reduces the block-copolymer free energy to Landau-Brazovskii form, the generic description that connects this model's phase behavior to other modulated-phase systems.","marker":"[93]"}],"fun_headline_variants":["Stiffness shrinks microphase domains and lowers ordering temperature","Domain size scales as stiffness^-1/2 in swollen elastomers","Nonlocal elasticity alone selects finite-size microdomains","Phase diagram links rubber stiffness to stripe and droplet patterns"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The entire stiffness dependence rests on one premise: the length scale below which the network feels no strain, $h$, is a fixed multiple $n \\approx 110$ of the end-to-end distance between crosslinks, $\\xi$, which shrinks with stiffness as $Y^{-1/2}$; if $h$ were instead a fixed material length, the predicted $\\Lambda \\sim Y^{-1/2}$ scaling and the linear drop of the ordering temperature with $Y$ would both disappear.","fun_headline_variants_meta":{"raw":{"variants":["Stiffness shrinks microphase domains and lowers ordering temperature","Domain size scales as stiffness^-1/2 in swollen elastomers","Nonlocal elasticity alone selects finite-size microdomains","Phase diagram links rubber stiffness to stripe and droplet patterns"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000336,"raw_usage":{"total_tokens":1872,"prompt_tokens":969,"completion_tokens":903,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":585,"completion_tokens_details":{"reasoning_tokens":835}},"tokens_in":585,"tokens_out":903,"duration_ms":9671,"temperature":1.0,"reasoning_tokens":835,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T20:13:44.775718+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"On the same PDMS samples, measure the structure-factor peak position $q_m$ by small-angle scattering during a temperature quench and independently determine the strand length $\\xi$ from swelling or rheology. The model requires the product $q_m\\xi$ to be a stiffness-independent constant equal to $\\sqrt{\\ln\\gamma}/n$ with $n = 110$, so $q_m$ must grow as $Y^{1/2}$, and requires $\\Lambda$ to keep following $Y^{-1/2}$ beyond the fitted 10–800 kPa range. A drift of $q_m\\xi$ with stiffness, or a saturation of $\\Lambda$ at extreme stiffnesses, would falsify the $h = n\\xi$ premise and the scaling laws built on it.","supporting_citations":[{"cited_title":"Rabin and A","cited_arxiv_id":null,"evidence_quote":"Defines the rescaled longitudinal modulus M of polymer gels that enters the elastic contribution to the free energy."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes the scaling of the strand end-to-end distance ξ with the dry Young's modulus Y, the source of the h ~ Y^(−1/2) dependence behind Λ ~ Y^(−1/2)."},{"cited_title":"Leibler and D","cited_arxiv_id":null,"evidence_quote":"Supplies the estimate of the interfacial parameter κ ~ k_BT/ℓ used to set γ and the absolute domain size."},{"cited_title":"Leibler, Theory of microphase separation in block copolymers, Macromolecules 13, 1602 (1980)","cited_arxiv_id":null,"evidence_quote":"Provides the block-copolymer microphase separation theory whose phase-diagram topology the model is shown to reproduce."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Reduces the block-copolymer free energy to Landau-Brazovskii form, the generic description that connects this model's phase behavior to other modulated-phase systems."}],"review_version":1}