Nanoconfinement Effects on Intermolecular Forces Observed via Dewetting

Evon S. Petek; Reika Katsumata; Tara (Tera) Huang

REVIEW 3 major objections 4 minor 4 references

Nanoconfinement Effects on Intermolecular Forces Observed via Dewetting

T0 review · 3 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Nanoconfinement in a PMMA underlayer shifts its refractive index enough to control whether a 10 nm polystyrene film dewets, and a modified Hamaker-constant model reproduces the measured dewetting wavelengths.

desk verdict The experimental observation is real and the new twist is a legitimate application of known refractive-index confinement data, but the quantitative match rests on an extrapolated and post-hoc-selected RI(d) curve, so the paper deserves review rather than acceptance as-is. read the letter →

arxiv 2506.03387 v1 pith:S4Z6G37Z submitted 2025-06-03 cond-mat.soft

Tara (Tera) Huang , Evon S. Petek , Reika Katsumata This is my paper

classification cond-mat.soft

keywords dewettingnanoconfinementHamakerconstantrefractiveindexvanderWaalsforcesthinpolymerfilmsfilmstabilityPS/PMMA/PStrilayer

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper proposes that when a polymer underlayer is squeezed to nanoscale thickness, its material properties change enough to alter the long-range van der Waals forces acting on a thin film above it. In a model trilayer—a 10 nm polystyrene (PS) film on a poly(methyl methacrylate) (PMMA) layer of thickness 15–95 nm on bulk PS—the dewetting wavelength of the top film increases as the PMMA gets thinner, meaning the film becomes more stable. Classical calculations using bulk refractive indices predict much less stabilization than observed. The authors trace the discrepancy to nanoconfinement: the PMMA refractive index depends on layer thickness, and feeding that thickness-dependent refractive index into the Hamaker constant reproduces the measured dewetting wavelengths. If correct, this makes underlayer thickness a knob for tuning thin-film stability without changing any chemistry.

What carries the argument

The load-bearing object is the effective Hamaker constant A23(d) for the interaction between PMMA (layer 2) and the top PS film (layer 3), computed through the Lifshitz formula using the PMMA refractive index as a function of thickness and then combined with PS via the pairwise approximation. This constant enters the linear-stability wavelength formula λ0 = 4π(πγ3)^(1/2) $h^{2}$ [(A11 − A31) + (1 + d/h)^(−2)(A31 − A21)]^(−1/2), which converts the modified intermolecular forces into the measurable dewetting pattern spacing. The refractive-index dependence is what carries the nanoconfinement effect: as d shrinks, n2(d) changes, A23 changes, and the predicted λ0(d) shifts toward the measured values.

What would settle it

Directly measure the refractive index of the 15 nm PMMA layer in this exact PS/PMMA/PS stack, for example by ellipsometry on a transferable trilayer or by combining neutron reflectometry density profiles with the Lorentz–Lorenz relation, and compare it with the Han extrapolated curve. If the true n2 at 15 nm differs enough that the effective A23 moves the predicted λ0 outside the experimental band in Fig. 3(c), the model's central claim fails. A cheaper check is to measure λ0 for d below 15 nm, where the polynomial predicts its steepest refractive-index change, and see whether the measured curve follows.

Watch

Extended reading notes

Core claim

The central claim is that nanoconfinement of a polymer underlayer changes its dielectric response enough to shift the intermolecular force landscape, and that this shift is visible in the characteristic wavelength λ0 of dewetting. Treating PMMA as a bulk material with refractive index n2 = 1.475 underpredicts the stability of a 10 nm PS film for PMMA thicknesses d ≲ 45 nm, whereas using a thickness-dependent refractive index n2(d) in the Lifshitz calculation of the Hamaker constant A23 brings theory into agreement with experiment. The paper therefore claims that the effective Hamaker constant, not the bulk value, is the right input for stability predictions in soft multilayer films, and that dewetting measurements provide direct evidence that nanoconfinement modifies long-range van der Waals interactions.

Load-bearing premise

The whole quantitative match rests on assuming that the PMMA refractive-index-versus-thickness curve measured by Han et al. on silicon-supported films, extended by a fourth-order polynomial from 35 nm down to 15 nm, also describes PMMA when it is sandwiched between PS layers, and that no other confinement effect matters at the same scale.

Editorial extensions

If this is right

A 10 nm polystyrene film can be stabilized or destabilized simply by choosing the PMMA underlayer thickness, with no change to the top film's chemistry or thickness.
Bulk Hamaker constants are not enough for polymer-on-polymer stability predictions in the nanoconfinement regime; the underlayer's thickness-dependent dielectric properties must be included.
Dewetting wavelength measurements can serve as a probe of nanoconfinement-induced optical changes in buried polymer underlayers.
Nanoconfinement effects already documented for glass-transition temperature, density, and refractive index should also enter predictions of long-range intermolecular forces in multilayer films.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

If the refractive-index mechanism is quantitative, the dewetting wavelength could be inverted to extract the nanoconfined refractive index of a buried underlayer without direct optical access; the paper does not make this inversion, but the monotonic λ0(d) curve makes it plausible.
The same correction should apply to other polymer pairs, but the direction of the effect depends on whether confinement raises or lowers the underlayer's refractive index, so the model predicts opposite stability trends for high-index versus low-index underlayers.
Because the analysis uses only early-stage, time-independent wavelengths, the nanoconfined Hamaker constants have not been tested against late-stage hole-growth kinetics; checking whether they also predict late-stage coarsening would extend the claim.
The paper leaves the 10 nm PS top layer's own nanoconfinement out of the calculation; including a thickness-dependent refractive index for the dewetting layer would be a natural refinement and would likely sharpen the quantitative match.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Short version: the paper has a genuine experimental result and a plausible explanation, but the numerical support for that explanation is weaker than the narrative implies because the thickness-dependent refractive index curve is extrapolated from a different system and chosen from among conflicting sources after the data were in hand. The paper is worth reading and worth refereeing, but the quantitative agreement should be treated as provisional, not a confirmed result.

What is actually new: the authors show that a 10 nm PS film dewetting on a PMMA underlayer has a characteristic wavelength that increases as the PMMA layer thins from 95 nm to 15 nm, and they interpret this as a nanoconfinement effect on the PMMA refractive index entering the Hamaker constant. I am not aware of anyone combining known thickness-dependent refractive index data with a Hamaker calculation to predict dewetting wavelengths in a polymer trilayer. That is a legitimate new application of existing physics, not a new mechanism. The experiments are careful: constant spin speed, water-transfer controls, OM/AFM cross-validation, and a two-stage dewetting analysis. The paper is also honest about its own limitations—interfacial width, competing RI datasets, and confinement effects in the top PS layer are all mentioned explicitly.

The soft spot is real and it is the load-bearing one. The solid line in Fig. 3(c) uses a fourth-order polynomial fit to Han et al.'s RI(d) data, measured on silicon-supported PMMA films thicker than 35 nm, extrapolated down to 15 nm in this PS/PMMA/PS stack. No measurement of n2(d) in this geometry or after the annealing history is reported. The paper shows two other published RI(d) trends that do not match Han's, but it does not compute lambda0(d) with those alternatives, so the reader cannot tell whether the agreement is a robust consequence of confinement or a post-hoc selection. On top of that, the calculation uses a single-wavelength refractive index at 632.8 nm to represent the full dielectric spectrum in a single-oscillator Lifshitz formula; that is a common approximation, but it is an extra layer of assumption. The experimental error bars also come from four measurements on one sample per thickness, so sample-to-sample variability is not captured.

Does the central argument hold up? Yes, in qualitative form. The bulk prediction clearly misses the trend at small d, and the confinement-corrected prediction does much better. The precise numerical match, however, is not established. The paper would be stronger if it showed lambda0(d) with uncertainty bands from the RI fit and at least one alternative RI dataset, or better yet, a direct ellipsometric measurement of PMMA refractive index in the trilayer geometry.

Who is this for? Researchers working on polymer thin film stability, dewetting, or nanoconfinement effects on optical properties. It deserves a serious referee, but the referee should be asked to focus on the RI extrapolation and the robustness of the quantitative fit.

My recommendation: send it to peer review. The idea is new enough and the experiments are solid enough. Require the RI-selection and extrapolation issue to be addressed before acceptance.

Referee Report

3 major / 4 minor

Summary. The paper reports dewetting experiments on a trilayer stack: a thick PS base, a PMMA middle layer of tunable thickness d = 15–95 nm, and a 10 nm top PS film. The authors measure the characteristic dewetting wavelength λ0 from optical microscopy and FFT analysis and find that λ0 increases as the PMMA layer becomes thinner, in qualitative agreement with a stabilizing underlayer effect but in quantitative disagreement with a bulk Hamaker-constant prediction at small d. They then replace the bulk PMMA refractive index with a thickness-dependent refractive index taken from Han et al. and recalculate the Hamaker constant via the Lifshitz equation, obtaining a modified theoretical curve that tracks the experimental λ0(d) more closely. The paper concludes that nanoconfinement-induced changes in the PMMA refractive index significantly modify long-range van der Waals interactions in polymer multilayers.

Significance. If the central claim holds, the paper makes a useful contribution: it connects thickness-dependent material properties (refractive index) to the macroscopic stability of polymer thin films and suggests a non-chemical route to tune film stability. The experimental work is careful and well controlled: the PMMA and PS control samples show that the observed instability is due to the PMMA middle layer, the two-stage dewetting analysis is time-resolved, and the theory is not fitted to the dewetting data. However, the quantitative agreement in Fig. 3(c) rests on an unmeasured, extrapolated refractive-index-versus-thickness curve for PMMA and on a specific choice among conflicting literature datasets. Because the supporting RI(d) input is load-bearing, the significance of the result for the broader nanoconfinement community is currently conditional rather than established.

major comments (3)

[SI Section F, Fig. S8; Results and Discussion ("Among several potential factors...")] The quantitative claim in Fig. 3(c) rests on an n2(d) curve that is not measured in the trilayer geometry of this paper. Han et al.'s data were obtained on PMMA supported on silicon at thicknesses ≥35 nm, and the fourth-order polynomial fit is extrapolated down to 15 nm, which is precisely the range where the bulk prediction fails. No ellipsometric measurement of n2(d) for the water-floated, annealed PMMA layer between PS layers is reported. Please provide a direct measurement (or an explicit quantitative argument for transferability) and propagate the polynomial fit uncertainty into the λ0 curve; without this, the agreement in Fig. 3(c) is not load-bearing evidence for nanoconfinement-modified Hamaker constants.
[SI Fig. S8; main text "Second, we recognize that the thickness dependence..."] The authors show in Fig. S8 that published PMMA refractive-index-thickness trends disagree (Han et al., Unni et al., and Todorov/Lalova), yet the theory curve in Fig. 3(c) is computed only with Han's data. Because the choice of dataset was made after observing the dewetting results, the robustness of the claimed agreement is untested. Please compute λ0 with at least one alternative RI(d) relation (or with upper/lower bounds consistent with Fig. S8) and discuss whether the nanoconfinement conclusion survives the choice; if it does not, the conclusion should be weakened.
[Eq. (S2) and Eq. (3) in main text] The Lifshitz calculation uses a single refractive index at 632.8 nm to represent the full dielectric response at the electronic absorption frequency via a single-oscillator approximation. If confinement changes the optical-frequency response differently from the relevant UV/static contributions, the inferred variation of A23(d) could be an artifact. Please assess the sensitivity of λ0(d) to the assumed absorption frequency νe and to the use of n at other wavelengths, or provide spectroscopic evidence that the confinement-induced change is broadband.

minor comments (4)

[Throughout (e.g., Eq. (3), Fig. 3 caption)] Please correct typos: "Plank" should be "Planck", "adsorption frequency" should be "absorption frequency", and "lower bond" should be "lower bound".
[Throughout (main text and figures)] The manuscript uses "l", "λ", "λ0", and "l0" interchangeably; please unify the notation for the characteristic wavelength.
[Fig. 3(a,b) and Methods] The error bars are stated to come from four measurements on a single sample per thickness; please add a comment on sample-to-sample reproducibility or explicitly acknowledge this limitation in the Methods.
[References 49/50 and SI Fig. S8] The text refers to "Todorov" while the SI cites "Lalova, A. & Todorov, R."; please use a consistent citation label for this dataset.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the model is a forward calculation from external literature refractive-index data, not fitted to the dewetting measurements.

full rationale

The paper's derivation chain combines Eq. (2) for the characteristic dewetting wavelength, Lifshitz theory (Eq. 3 / Eq. S2), and a pairwise approximation (Eq. S3). The only non-standard input is the PMMA refractive-index-thickness relation, which is taken from the independent published work of Han et al. (ref 38) and extrapolated in the SI. No parameter is fit to the experimental lambda_0 values: the solid curve in Fig. 3(c) is a forward computation using literature RI(d) and textbook Lifshitz theory. The choice of Han et al. over Unni et al. or Todorov et al. is a data-selection and extrapolation limitation, and the paper explicitly acknowledges conflicting RI trends, but this does not make the prediction equivalent to the experimental output by construction. Self-citations (refs 9, 40) appear in the motivation and discussion, not as load-bearing steps in the derivation. Therefore no circular step can be exhibited under the required standard, and the appropriate finding is no significant circularity.

Assumptions & free parameters 1 free parameters · 5 assumptions · 0 invented entities

The model has no free parameters fitted to the dewetting data, but it relies on an externally fitted RI(d) relation, extrapolated below the measured thickness range, and on the transferability of that relation to a different substrate architecture. The central claim therefore rests on literature data and several simplifying assumptions about which material property changes dominate.

free parameters (1)

Fourth-order polynomial coefficients for RI(d) fit = Not reported
The nanoconfined Hamaker constant A23(d) and hence the predicted lambda0 depend on this fit to Han et al.'s refractive index data. The coefficients are not given in the paper or SI, so the exact prediction cannot be reproduced without redigitizing Fig. S8.

assumptions (5)

domain assumption Lifshitz equation (Eq. 3) with a single electronic absorption frequency ve = 3e15 Hz adequately describes Hamaker constants for polymers.
Standard approximation in Israelachvili; uses a typical frequency rather than measured values for PS/PMMA. Affects absolute Hamaker values but not the thickness dependence unless RI changes.
standard math Eq. 2 (Sharma-Reiter) correctly predicts characteristic wavelength lambda0 from Hamaker constants and surface tension.
Prior literature model for spinodal dewetting of thin films on coated substrates; accepted in the field.
domain assumption The PMMA middle layer in the PS/PMMA/PS trilayer exhibits the same RI-thickness relationship as PMMA films on silicon in Han et al.
Transferability of confinement-induced RI changes across different substrates is assumed; no direct ellipsometry on the trilayer is presented.
ad hoc to paper Only the PMMA refractive index changes with confinement; the PS layers and PS/PMMA interface retain bulk properties.
Authors acknowledge interfacial width and PS nanoconfinement as potential factors, but omit them from the model, stating they are small or difficult to measure. This assumption is necessary for the simplified prediction.
domain assumption The silicon substrate's van der Waals contribution is negligible because the bottom PS layer is >400 nm thick.
Stated in the text: vdW interactions from silicon are significant only for PS film thicknesses below 20 nm, much smaller than the 400 nm bottom layer.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Nanoconfinement Effects on Intermolecular Forces Observed via Dewetting." pith.science (2026). https://pith.science/paper/S4Z6G37Z

@misc{pith2026250603387,
  author       = {Pith},
  title        = {Pith review of: Nanoconfinement Effects on Intermolecular Forces Observed via Dewetting},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/S4Z6G37Z}},
  note         = {Machine review of arXiv:2506.03387}
}

read the original abstract

Although wettability is a macroscopic manifestation of molecular-level forces, such as van der Waals (vdW) forces, the impact of nanoconfinement on material properties in reduced film thickness remains unexplored in predicting film stability. In this work, we investigate how nanoconfinement influences intermolecular interactions using a model trilayer system composed of a thick polystyrene (PS) base, a poly(methyl methacrylate) (PMMA) middle layer with tunable thickness (15-95 nm), and a 10 nm top PS film. We find that the dewetting behavior of the top PS layer is highly sensitive to middle PMMA thickness, deviating from classical vdW-based predictions that assume bulk material properties. By incorporating nanoconfinement-induced changes in PMMA refractive index into the calculation of the Hamaker constant, we present a modified theoretical framework that successfully captures the observed behavior. This study links dewetting behavior and material property change as a function of underlayer thickness, providing direct evidence that nanoconfinement in soft matter systems significantly influences long-range intermolecular interactions. We show that film stability can be tuned solely by adjusting underlying layer thickness, while preserving both chemistry and thickness of top functional film. This finding carries broad implications for thin-film technologies across scientific and engineering disciplines by enabling performance-targeted interface design.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

4 extracted references · 4 canonical work pages

[2]

Multilayer polymer system The refractive index data (RI, open circles) was sourced from Han et al.’s work, and data points for thicknesses less than 150 nm were fitted with a 4th order polynomial to generate full RI curves within this range (Fig. S8)5. Since the thinnest PMMA thickness in Han et al.’s experiments was 35 nm, we first fitted the data from 3...

work page 2001
[24]

Wyart, F. B. & Daillant, J. Drying of solids wetted by thin liquid films. Can. J. Phys. 68, 1084–1088 (1990). 25. Israelachvili, J. N. Intermolecular and Surface Forces. Academic Press, Cambridge (1985). 26. Maldarelli, C., Jain, R. K., Ivanov, I. B. & Ruckenstein, E. Stability of symmetric and unsymmetric thin liquid films to short and long wavelength pe...

work page 1990
[37]

& Kimura, S

Liang, T., Makita, Y. & Kimura, S. Effect of film thickness on the electrical properties of polyimide thin films. Polymer 42, 4867–4872 (2001). 38. Han, Y., Huang, X., Rohrbach, A. C. W. & Roth, C. B. Comparing refractive index and density changes with decreasing film thickness in thin supported films across different polymers. The Journal of Chemical Phy...

work page 2001
[45]

#(μm-1) log t (hr) 2

Xu, L. et al. Dewetting kinetics of thin polymer bilayers: Role of under layer. Polymer 52, 4345–4354 (2011). 46. Xie, F. et al. Distinguishing Spinodal and Nucleation Phase Separation in Dewetting Polymer Films. in Nano Science and Technology (eds. Sheng, P. & Tang, Z.) 184–189 (CRC Press, 2003). doi:10.1201/9780203390283.ch19. 47. Williams, M. B. & Davi...

work page doi:10.1201/9780203390283.ch19 2011

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.

[1] [2]

Multilayer polymer system The refractive index data (RI, open circles) was sourced from Han et al.’s work, and data points for thicknesses less than 150 nm were fitted with a 4th order polynomial to generate full RI curves within this range (Fig. S8)5. Since the thinnest PMMA thickness in Han et al.’s experiments was 35 nm, we first fitted the data from 3...

work page 2001

[2] [24]

Wyart, F. B. & Daillant, J. Drying of solids wetted by thin liquid films. Can. J. Phys. 68, 1084–1088 (1990). 25. Israelachvili, J. N. Intermolecular and Surface Forces. Academic Press, Cambridge (1985). 26. Maldarelli, C., Jain, R. K., Ivanov, I. B. & Ruckenstein, E. Stability of symmetric and unsymmetric thin liquid films to short and long wavelength pe...

work page 1990

[3] [37]

& Kimura, S

Liang, T., Makita, Y. & Kimura, S. Effect of film thickness on the electrical properties of polyimide thin films. Polymer 42, 4867–4872 (2001). 38. Han, Y., Huang, X., Rohrbach, A. C. W. & Roth, C. B. Comparing refractive index and density changes with decreasing film thickness in thin supported films across different polymers. The Journal of Chemical Phy...

work page 2001

[4] [45]

#(μm-1) log t (hr) 2

Xu, L. et al. Dewetting kinetics of thin polymer bilayers: Role of under layer. Polymer 52, 4345–4354 (2011). 46. Xie, F. et al. Distinguishing Spinodal and Nucleation Phase Separation in Dewetting Polymer Films. in Nano Science and Technology (eds. Sheng, P. & Tang, Z.) 184–189 (CRC Press, 2003). doi:10.1201/9780203390283.ch19. 47. Williams, M. B. & Davi...

work page doi:10.1201/9780203390283.ch19 2011