{"id":"b4c50223-b2c9-4f0d-8370-218201916716","arxiv_id":"2608.09375","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"A new FMT-style weighted-density functional for square-well mixtures predicts that liquid-vapor interfaces attract nano particles via an effective one-body potential, computed consistently by density-profile and Widom insertion routes.","lead":"Using a new weighted-density form of the random phase approximation for square-well mixtures, the authors compute the effective one-body potential that a liquid-vapor interface exerts on dissolved nano particles. The potential strongly attracts particles to the interface, and two analytical routes to compute it agree, though the agreement is built into the density functional method.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Eq. (17) is not equivalent to the RPA functional Eq. (9) for non-uniform densities: the Fourier-space kernels differ at finite wavevectors, so the interface profiles and V_eff are computed with a different, unvalidated perturbation functional.","rationale":"The reader identified the accuracy of the RPA perturbation term as the weakest point and issued a conditional accept. My stress-test found a more basic and more damaging issue: the paper's own formal claim that Eq. (17) is equivalent to Eq. (9) does not hold for non-uniform density profiles. The bulk free energy match (Eq. 20 versus Eq. 12) is real, and the two routes to V_eff (Eqs. 35 and 36) are equivalent within the implemented functional, so the internal consistency is not in question. But the second functional derivative of Eq. (17) differs from the Fourier transform of the square-well kernel of Eq. (9) at finite wavevector. Since the interface profile and V_eff(z) are determined by exactly those finite-wavevector properties, the numerical results are not a realization of the RPA perturbation theory the paper claims to use. The 'excellent agreement' between the two routes is a check on the numerical solution of the new functional, not a validation of it. This does not disprove the physical conclusion that this approximate functional attracts nano particles to the interface, but it invalidates the central theoretical claim as written. The paper could potentially be revised to present Eq. (17) as a new approximation that reduces to RPA only in bulk, with its accuracy benchmarked against simulation; but in its current form, the central claim is unsupported. I therefore recommend rejecting or, at minimum, requiring a major revision that either proves the equivalence or abandons it.","tokens_in":14599,"tokens_out":28127,"duration_ms":244288,"concrete_test":"Evaluate analytically the ratio of the two Fourier kernels at ka=1, with a = λσ/2: K_new(1)/K_RPA(1) = [j0(1)j1(1) + j0(1)^2 + j1(1)^2] / [4 j1(2)/2] ≈ 1.64. If the ratio is not exactly 1, then Eq. (17) is not the RPA functional Eq. (9), and the manuscript's central formal claim fails for non-uniform density profiles.","verdict_should_be":"REJECT","load_bearing_attack":"The formal foundation of the paper is the claim in Sec. 2.2 that the weighted-density expression Eq. (17) is 'a functional equivalent to Eq. (9)'. This claim fails for inhomogeneous densities. For a one-component SW fluid, with a = λR (R = σ/2 so the RPA range is 2a), the second functional derivative of Eq. (17) has Fourier kernel K_new(k) = -4π ε a^3 [j0(ka)j1(ka)/(ka) + j0(ka)^2 + j1(ka)^2], whereas Eq. (9) with the potential of Eq. (10) gives K_RPA(k) = -(1/2) ε ∫_{r<2a} e^{-ik·r} d^3r = -16π ε a^3 j1(2ka)/(2ka). Both kernels have the same k=0 limit, -16π/3 ε a^3, which is why the bulk free energy, Eq. (20), matches Eq. (12). At finite k they differ; e.g. at ka=1 the ratio is about 1.64. Thus the identity asserted after Eq. (16) holds only in the bulk, not as a density functional. The interface calculation is a legitimate DFT calculation of a different, FMT-inspired perturbation term, but it is not the RPA calculation described in Sec. 2.1. The agreement between Eqs. (35) and (36) is an internal consistency check of that different functional and does not validate the RPA mapping. The central claim of the paper, as stated, is therefore not supported.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript proposes a weighted-density reformulation of the square-well perturbation contribution to classical density functional theory (Eq. 17), claims that this form is equivalent to the standard optimized random-phase approximation (RPA) functional (Eq. 9), and applies the new functional to an additive binary square-well mixture at liquid-vapor coexistence. The authors compute density profiles of a dilute nanoparticle component, report a reduction of the interfacial surface tension, and obtain an effective one-body interface potential V_eff(z) by two routes: from the solute density profile (Eq. 35) and from the Widom insertion expression (Eq. 36). The central physical claim is that the liquid-vapor interface attracts nanoparticles strongly, with density enhancements reaching five orders of magnitude for size ratio sigma_2/sigma_1 = 3.","tokens_in":14940,"tokens_out":13272,"duration_ms":113137,"significance":"If correct, the paper would deliver a computationally convenient FMT-style functional for multicomponent square-well fluids and a clean DFT route to interface-induced one-body effective potentials. The manuscript has genuine strengths: the bulk limit of the new functional reproduces the RPA free energy under Lorentz-Berthelot mixing; the Widom insertion route is elegant and allows the dilute limit to be taken analytically; and the numerical implementation is transparent. However, the claimed equivalence with RPA is not established for inhomogeneous densities, and the apparent validation in Fig. 4 is a formal identity rather than an independent check. As a result, the quantitative predictions, including the large density enhancement and the surface-tension reduction, are supported only by an unbenchmarked functional.","major_comments":[{"comment":"The statement that Eq. (17) is 'the RPA of the SW fluid, now in the fashion of FMT' is not correct for non-uniform densities. For a one-component fluid with a = lambda R, the second functional derivative of Eq. (17) has the Fourier kernel K_new(k) = -4 pi epsilon a^3 [j0(ka)j1(ka)/(ka) + j0(ka)^2 + j1(ka)^2], while the RPA functional (9) with the potential (10) has K_RPA(k) = -4 pi epsilon (2a)^3 j1(2ka)/(2ka). The two kernels agree only at k = 0, which is why the bulk free energy (20) matches (12); at finite k they differ, with the magnitude ratio around 1.65 at ka = 1. Since the interface profile is intrinsically inhomogeneous, the calculations in Sec. 3 use a different perturbation functional from the RPA described in Sec. 2.1. The paper must either prove the functional identity for arbitrary densities or explicitly reframe Eq. (17) as an FMT-inspired approximation and benchmark it against Eq. (9) in the planar-interface geometry.","section":"Sec. 2.2, Eq. (17)"},{"comment":"The 'excellent agreement' between the density-profile route and the Widom-insertion route is a formal consequence of the DFT Euler-Lagrange equation, not an independent validation. Substituting Eq. (7) into Eq. (35) gives beta V_eff(z) = c_2^(1)(infinity) - c_2^(1)(z), which is exactly Eq. (36). Thus Fig. 4 primarily checks the numerical consistency of the solver, and the paper should state this explicitly rather than presenting the agreement as evidence for the accuracy of V_eff.","section":"Sec. 3.2, Eqs. (35)-(36)"},{"comment":"No external benchmark is provided for the RPA-based perturbation term, which controls the phase diagram, the interface structure, the surface tension, and hence V_eff. The manuscript cites Archer et al. for the general accuracy of RPA but does not test the square-well solvent at coexistence against simulation or experiment. Given that the quantitative claims, such as the five-orders-of-magnitude density enhancement in Fig. 3(d) and the surface-tension reductions in Table 1, are sensitive to this term, a comparison with simulation or with the full RPA functional (9) for at least one state point is needed before those magnitudes can be accepted.","section":"Sec. 2.1 and Sec. 3"}],"minor_comments":[{"comment":"In the Summary and Outlook, 'Wigner's insertion theorem' should be 'Widom's insertion theorem'.","section":"Sec. 4"},{"comment":"The same symbol w^{epsilon,i}_2 is used for both the scalar weight and the vector weight; please distinguish the vector weight with boldface or an arrow.","section":"Eq. (16)"},{"comment":"The DFT and Gibbs columns differ by roughly a factor of four at sigma_2/sigma_1 = 3, so 'good agreement' overstates the match; the discrepancy should be discussed or the wording softened.","section":"Table 1"},{"comment":"The sentence about including rho_2(z -> infinity) = rho_{2,0} as a 'fourth equation to Eq. (25)' is confusing because rho_{2,0} is already fixed when solving the coexistence conditions; please clarify the numerical procedure.","section":"Sec. 3.1, around Eq. (25)"},{"comment":"The sigmoid parameter a is obtained from a fit, but the fit range and quality are not reported; since Eq. (29) is used in the Gibbs adsorption analysis, please provide this information.","section":"Sec. 3.1, Eq. (28)"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is clearly written and the technical execution appears careful, but the central equivalence claim between Eq. (17) and the RPA functional is not supportable as stated. The authors should either demonstrate the functional identity for inhomogeneous densities or substantially reframe the contribution as an FMT-inspired approximation and validate it against the actual RPA functional or simulation. Because the proposed weighted-density form may still be useful, revision is appropriate rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe one thing you should know: the paper's central claim, that the FMT-style weighted-density functional Eq. (17) is equivalent to the standard RPA functional Eq. (9), is false for inhomogeneous densities. I worked out the second functional derivatives. RPA gives a kernel whose Fourier transform is essentially j1(2ka)/(2ka). Eq. (17) gives a kernel built from j0(ka)j1(ka)/(ka) + j0^2 + j1^2. The two agree at k=0, which is why the bulk free energy matches, but they differ at finite wavevector; the ratio at ka=1 is around 1.2, not the 1.64 the stress-test quote suggests, but the difference is real. So the paper is not a reformulation of RPA. It is a new, unvalidated perturbation functional. The interface profiles and effective potentials are computed with this new functional, not with the RPA the authors cite.\n\nWhat is genuinely new and useful: the FMT-style weighted-density representation of the square-well attraction is a clever construction, and it may indeed be convenient for multi-component mixtures. The application to the effective one-body potential at a liquid-vapor interface, with two independent computational routes (density profile and Widom insertion) agreeing, is a nice result. But that agreement is a self-consistency check, not a validation: both expressions reduce to the same direct-correlation difference via the Euler-Lagrange equation.\n\nSoft spots: no simulation or experimental comparison anywhere, so the large adsorption enhancements (up to five orders of magnitude) are entirely uncalibrated. The sigmoidal fit parameter a is ad hoc. The surface-tension reduction is tiny and not validated.\n\nThe paper is not sound as written because its main identity fails in the inhomogeneous case. But the idea has merit. I'd send it to a serious referee with the expectation of major revision: either prove the equivalence (I don't think it's true) or drop the RPA claim, present Eq. (17) as a new approximation, and benchmark it against simulation.","headline":"The paper's central equivalence between its new weighted-density functional and RPA fails for inhomogeneous densities, so the interface results rest on an unvalidated approximation, but the idea is worth a careful revision.","tokens_in":15475,"tokens_out":21403,"would_cite":false,"duration_ms":161787,"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":"A liquid-vapor interface of a square-well solvent exerts a strong attractive one-body force on dilute nanoparticles, and a new weighted-density functional gives two agreeing routes to it.","keywords":["classical density functional theory","square-well mixture","weighted densities","fundamental measure theory","random phase approximation","effective interaction","liquid-vapor interface","surface tension"],"falsifier":"A Monte Carlo or molecular-dynamics simulation of the same additive binary square-well mixture ($\\beta\\varepsilon_1=1$, $\\lambda_1=1.5$, $\\sigma_2/\\sigma_1$ from 1 to 3, with the symmetric and solvophilic solute energies used in Sec. 3) could measure the solute density profile across the planar liquid-vapor interface and the surface tension; if the interface accumulation at the largest size ratio does not approach five orders of magnitude, or the surface-tension reduction disagrees with the values in Table 1, the RPA solvent model underlying $V_{\\mathrm{eff}}(z)$ would be falsified.","tokens_in":14377,"feed_emoji":"🧲","tokens_out":14962,"duration_ms":115958,"temperature":0.7,"pith_summary":"Classical density functional theory for square-well fluid mixtures usually treats the attractive part with the random-phase approximation, which is accurate but cumbersome to set up for many components. This paper rewrites the square-well perturbation functional in the weighted-density form of fundamental measure theory, Eq. (17), recovering the random-phase approximation exactly for additive mixtures with Lorentz-Berthelot mixing rules. Applying this functional to a binary square-well mixture at liquid-vapor coexistence, the paper finds that a dilute solute of nano particles is strongly attracted to the interface, with equilibrium density enhancements growing up to five orders of magnitude at the largest size ratio studied. It then establishes the paper's central quantitative claim: the effective one-body potential acting on the nano particles can be computed either from the solute density profile, Eq. (35), or from the Widom insertion theorem, Eq. (36), and the two routes agree closely. If correct, this gives a simple way to compute interface-induced forces on dilute solutes and the accompanying reduction of surface tension.","feed_headline":"New functional predicts nanoparticles flock to liquid-vapor interfaces","feed_subtitle":"At size ratio three, nanoparticle density at the interface jumps five orders of magnitude.","key_machinery":"The load-bearing object is the new weighted-density form of the square-well perturbation functional, Eq. (17), built from effective FMT-style weight functions $w_3^{\\varepsilon,i}$, $w_2^{\\varepsilon,i}$, $w_1^{\\varepsilon,i}$, $w_0^{\\varepsilon,i}$, and vector weights that carry factors of $\\sqrt{\\varepsilon_i}$ and are evaluated at the square-well range $\\lambda_i R_i$. This form is an exact rewriting of the optimized random-phase approximation for additive square-well mixtures, so it keeps the full RPA thermodynamics while reducing the attraction term to the same convolution structure as the hard-sphere weighted densities. The second piece of machinery is the pair of formulas for the effective one-body potential, Eq. (35) from the dilute-solute density profile and Eq. (36) from Widom insertion; their agreement is the paper's main quantitative check on the mapping.","core_discovery":"The central claim is that the square-well attraction term of the excess free-energy functional, conventionally a double convolution in the optimized random-phase approximation, can be written exactly as a fundamental-measure-theory-style functional of weighted densities: $\\beta F_{\\mathrm{sw}}=\\int d\\mathbf{r}\\,\\beta(n_0^\\varepsilon n_3^\\varepsilon+n_1^\\varepsilon n_2^\\varepsilon-\\mathbf{n}_1^\\varepsilon\\cdot\\mathbf{n}_2^\\varepsilon)$, with effective weights built from the square-well energy and range of each species. For an additive binary square-well mixture at liquid-vapor coexistence, this functional predicts that the interface exerts on a dilute solute of nano particles an effective one-body potential $V_{\\mathrm{eff}}(z)$ that is strongly attractive at the interface. The paper shows that $V_{\\mathrm{eff}}(z)$ can be obtained either from the solute density profile through $\\beta V_{\\mathrm{eff}}(z)=-\\lim_{\\rho_{2,0}\\to 0}\\log(\\rho_2(z)/\\rho_{2,0})$ or from the change in the solute's one-body direct correlation function through $\\beta V_{\\mathrm{eff}}(z)=\\lim_{\\rho_{2,0}\\to 0}[c_2^{(1)}(\\pm\\infty)-c_2^{(1)}(z)]$, and it reports that both routes agree closely. The same calculations show that the interfacial accumulation lowers the surface tension, consistent with the Gibbs adsorption theorem.","pith_inferences":["Going beyond the paper, the same weighted-density construction should extend directly to square-shoulder attractions (the paper notes the sign of the weights would flip) and to ternary or polydisperse square-well mixtures, since the convolution structure is no more costly than the hard-sphere terms.","The predicted interfacial accumulation is strong enough that the derived $V_{\\mathrm{eff}}(z)$ could be used as an input to Brownian dynamics of drying droplets, making the coffee-stain transport mechanism quantitative in this model.","A direct test of the weakest point would be to compare the one-component square-well interface's density profile and surface tension against simulation; if the mean-field attraction model is inaccurate, the depth and sign of the predicted effective potential are the quantities most likely to change."],"forward_implications":["The weighted-density functional recovers the random-phase approximation for any number of additive square-well components, so phase equilibria, density profiles, and surface tensions of multi-component square-well mixtures can be computed with the same convolution structure as hard-sphere fundamental measure theory.","At the liquid-vapor interface of the square-well solvent, dilute nano particles experience a strongly attractive one-body potential, with the equilibrium density enhancement growing from an order of magnitude at $\\sigma_2/\\sigma_1=1.5$ to five orders of magnitude at $\\sigma_2/\\sigma_1=3$.","The Widom-insertion route to the effective potential, Eq. (36), requires only the pure solvent interface profile, so the dilute-limit one-body force can be obtained without solving the full binary mixture.","The interfacial accumulation lowers the surface tension, and the reduction is consistent with the Gibbs adsorption theorem, growing with the size ratio and reaching about 0.6% of the bare solvent value in the dilute regime studied."],"supporting_citations":[{"why":"Supplies the standard optimized random-phase approximation for attractive interactions, Eq. (9), that the new functional reproduces.","marker":"Hansen and McDonald (2006)"},{"why":"Argues that the standard mean-field treatment of attraction in classical DFT is better than expected, justifying the use of RPA for the solvent.","marker":"Archer et al. (2017)"},{"why":"Introduces fundamental measure theory and the low-density weighted-density form that the new square-well functional is built to imitate.","marker":"Rosenfeld (1989)"},{"why":"Provides the White-Bear fundamental measure theory used as the hard-sphere reference functional for all profiles and coexistence calculations.","marker":"Roth et al. (2002)"},{"why":"Supplies the BMCSL equation of state used for the hard-sphere mixture pressure in the bulk coexistence equations.","marker":"Mansoori et al. (2003)"},{"why":"States the insertion theorem that gives the excess chemical potential route to the effective one-body potential, Eq. (36).","marker":"Widom (1963)"},{"why":"Used alongside Widom's theorem as the basis for computing the change in grand potential on inserting a particle.","marker":"Henderson (1983)"},{"why":"Supplies the remove-and-insert procedure for effective interactions in inhomogeneous fluids that the Widom route applies to the interface.","marker":"Roth et al. (2000)"},{"why":"Provides an equivalent earlier investigation of local solubility at liquid-vapor interfaces with which the paper's density-profile analysis is compared.","marker":"Abe and Koga (2014)"}],"fun_headline_variants":["New functional shows nanoparticles love liquid-vapor interfaces","Interface attraction: square-well functional predicts nanoparticle pull","Nanoparticles drawn to vapor interface per new density functional","Effective one-body force: why nanoparticles migrate to interfaces","Theory: liquid-vapor interface exerts strong pull on nanoparticles"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that treating the attractive forces among solvent molecules as a mean-field average, without accounting for correlations between those attractions, is accurate enough at liquid-vapor coexistence; the paper tests this against no simulation or experiment.","fun_headline_variants_meta":{"raw":{"variants":["New functional shows nanoparticles love liquid-vapor interfaces","Interface attraction: square-well functional predicts nanoparticle pull","Nanoparticles drawn to vapor interface per new density functional","Effective one-body force: why nanoparticles migrate to interfaces","Theory: liquid-vapor interface exerts strong pull on nanoparticles"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000706,"raw_usage":{"total_tokens":3194,"prompt_tokens":971,"completion_tokens":2223,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":587,"completion_tokens_details":{"reasoning_tokens":2147}},"tokens_in":587,"tokens_out":2223,"duration_ms":16488,"temperature":1.0,"reasoning_tokens":2147,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T18:28:43.772366+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A Monte Carlo or molecular-dynamics simulation of the same additive binary square-well mixture ($\\beta\\varepsilon_1=1$, $\\lambda_1=1.5$, $\\sigma_2/\\sigma_1$ from 1 to 3, with the symmetric and solvophilic solute energies used in Sec. 3) could measure the solute density profile across the planar liquid-vapor interface and the surface tension; if the interface accumulation at the largest size ratio does not approach five orders of magnitude, or the surface-tension reduction disagrees with the values in Table 1, the RPA solvent model underlying $V_{\\mathrm{eff}}(z)$ would be falsified.","supporting_citations":[],"review_version":1}