{"id":"dd3a4e1f-e276-49fb-afb4-228f04bcf56e","arxiv_id":"1908.01305","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Water droplets self-propel along surfaces that are textured with silicone-oil-covered rails, because a liquid-on-liquid interface removes solid friction and the rail geometry provides a wettability gradient.","lead":"This paper creates a 'shaped liquid surface' by coating micro-scale rails with a thin layer of silicone oil, so water droplets glide on oil rather than touching solid. The design produces strong wettability gradients and near-zero friction, driving droplets over long paths, uphill, and even upside down.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Oil-filled air pockets at fs>0.7 (SI §1.4) break the Cassie-state assumption used for the driving force on the high-fs end of the gradient; the v∝1/fs model and full-pattern long-range claim are untested there.","rationale":"Reader and I converge on the same regime: the oil-filling of air pockets at high fs. I partially disagree with the reader's phrasing that 'solid-free contact could fail locally': oil-filled pockets still provide a liquid contact; the real casualty is the Cassie-state gradient. Because Eq. 2 is used both to predict the apparent-angle variation and to derive Fd, the filled state undermines the quantitative force balance at the high-fs destination. The paper's own SI provides the evidence, so this is not an external objection. The qualitative demonstration—low-hysteresis liquid surface, direction toward higher fs, uphill and inverted capture—is independently supported by movies and controls, so the verdict should remain CONDITIONAL rather than reject. I recommend no change to the reader's verdict; the conditional should explicitly require the high-fs transport check.","tokens_in":10138,"tokens_out":17054,"duration_ms":193613,"concrete_test":"Recover the raw position-vs-time tracks used in Fig. 2b; convert the x-axis to fs and plot v(fs) for each droplet volume. Overlay fs=0.7 and fit v∝1/fs only for fs<0.7; report residuals and the terminal position of each droplet. If droplets stop at fs≈0.7–0.8 with contact angle θe, the driving-force model must be restricted to the Cassie regime and the long-range claim re-scoped. If droplets continue to fs≈1 with v following 1/fs, the oil-filling observation does not affect transport.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central mechanism requires that the droplet remain in the composite liquid-air Cassie state described by Eq. (2) over its whole footprint, so that locally cosθ=-1+(1+cosθe)fl with fl=fs+0.059. This is the basis for the integrated driving force Fd=8kγoaαR^2 in SI §1.7.1 and for the quantitative comparison with the critical-angle measurements in Fig. 3c. The paper's SI §1.4 explicitly reports that for fs>0.7 the air pockets under the droplet are sometimes filled with oil, in which case the apparent contact angle falls to the flat-liquid value θe and loses its dependence on fs. On the transport substrate fs rises linearly from 0 to 1, so a droplet with base radius R must sample fs>0.7 before the end of the pattern. Once part of the droplet's footprint is in the filled state, d(cosθ)/dx is locally zero; if the entire footprint is filled, the integrated wettability-gradient force is zero and the droplet should stop before reaching the fs→1 end. The model therefore overestimates the driving force in the terminal regime, and the 'sustained over the pattern' or 'fs=0 to 1' transport claim is not supported by the presented data. Importantly, the failure mode is loss of wettability gradient, not loss of liquid contact: the filled state is still an oil surface, so the liquid-on-liquid aspect survives. The low-fs self-propulsion and scaling remain credible, but the long-range endpoint needs re-scoping or additional evidence.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript presents a dual-length-scale 'shaped liquid surface' formed by nanotextured rails imbibed with silicone oil. It reports measurements of low contact angle hysteresis, a linear variation of apparent contact angle with solid fraction, spontaneous droplet motion toward higher solid fraction, self-propelled transport along gradient rails with speed decreasing as 1/f_s, uphill and inverted transport, and capture of impacting droplets. A force-balance model balances a capillary driving force F_d = 8kγ_oa α R^2 with viscous dissipation at the contact line to yield v ∝ γ_oa α R/(μ_o f_s). The qualitative phenomena are directly demonstrated, while the quantitative model contains fitted corrections and unmeasured parameters.","tokens_in":10555,"tokens_out":6356,"duration_ms":66052,"significance":"If validated, the approach offers a passive, energy-free route to droplet transport with low hysteresis, including in inverted orientations, with potential applications in microfluidics, self-cleaning surfaces, fog harvesting, and heat transfer. The strengths are the clear experimental demonstrations, the use of control super-hydrophobic surfaces that remain pinned, the direct measurement of low hysteresis, and the explicit reporting of a limitation in SI §1.4. The main weaknesses are that the driving-force model relies on a fitted wetting law, an unreported footprint aspect ratio, and a Cassie-state assumption that is known to break down at f_s > 0.7; these issues make the quantitative comparison partly a consistency check rather than a falsifiable prediction.","major_comments":[{"comment":"SI §1.4 explicitly reports that for f_s > 0.7 the air pockets under the droplet are sometimes filled with oil, in which case the apparent contact angle falls to θ_e and loses its dependence on f_s. Because the gradient substrate spans f_s = 0 to 1, a droplet with a finite base radius must sample f_s > 0.7 before the end of the pattern. The driving force F_d = 8kγ_oa α R^2 in SI §1.7.1 is derived by integrating Eq. (2) over the entire footprint, so it overestimates the wettability contrast in the terminal portion of the transport. The 'sustained self-propulsion over the pattern' claim is therefore not supported by the presented data, which end near f_s = 0.77 in Fig. 2a; the authors should either restrict the claim to the regime where the composite Cassie state is stable, demonstrate that the droplet completes the pattern before entering the oil-filled regime, or include a model for the transition to the filled state.","section":"SI §1.4 and main text Fig. 2a"},{"comment":"The driving force expression F_d = 8kγ_oa α R^2 contains the droplet footprint aspect ratio k, but k is never measured or reported anywhere in the manuscript. The comparison in Fig. 3c between the measured driving force and the model is therefore not fully quantitative: an order-one value of k could be chosen to bring the model into agreement, and the reported error bars do not constrain k. To make the force-balance model testable, the authors should report k from direct footprint measurements or perform the integration over the actual contact-line shape.","section":"SI §1.7.1, Eq. (5)"},{"comment":"Eq. (2) is fit to the contact-angle data using the correction f_l = f_s + 0.059, and this fitted correction is inherited by the force-balance model, since SI §1.7.1 uses Eq. (2) directly. Consequently the quantitative agreement in Fig. 3c is not a parameter-free test of the driving-force model. The authors should clearly separate fitted from predicted quantities and show the sensitivity of F_d and the terminal velocity v to the fitted offset δ = 0.059, especially because the offset is close to the coating-thickness estimate and may vary between samples.","section":"Fig. 1c and Eq. (2)"},{"comment":"The measured pinning force F_p ≈ 1.11 ± 0.25 μN is more than a factor of two larger than the value estimated from contact angle hysteresis, F_p ≈ 0.48 μN. This discrepancy is not discussed, yet it is relevant to the claim that pinning is overcome by the gradient. The authors should explain whether the hysteresis-based estimate underestimates the relevant pinning or whether an additional dissipative or pinning mechanism contributes to the critical-angle measurements.","section":"Main text, paragraph following Fig. 3"}],"minor_comments":[{"comment":"The vertical axis label 'F [ N]' appears to be missing the micro symbol; it should read 'F [μN]' to match the values in the text.","section":"Fig. 3c axis label"},{"comment":"In Eq. (13) the symbol 'ν' is used where the droplet velocity 'v' is meant; this typo makes the equation momentarily confusing.","section":"SI §1.7.4, Eq. (13)"},{"comment":"The phrase 'the liquid surface area fractions f_l increases linearly' should be 'the liquid surface area fraction f_l increases linearly'.","section":"Main text, first paragraph of the propulsion section"},{"comment":"The sentence fragment 'the conical shape of cactus spines to create self-propelled motion' is grammatically incomplete and should be revised for clarity.","section":"Abstract"}],"recommendation":"major_revision","confidential_remarks":"The qualitative experimental results are convincing and the low-hysteresis liquid-on-liquid concept is a genuine contribution. The main concern is the high-f_s regime: the paper's own SI acknowledges the Cassie-state breakdown above f_s ≈ 0.7, yet the long-range transport claim and the quantitative model extend into that range. This is fixable by re-scoping the claims and providing data or a model for the oil-filled transition. The unreported aspect ratio k also weakens the quantitative comparison. I see no reason to reject, but the central quantitative claim needs revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The thing to know: this is a genuine experimental advance, not just another lubricant-infused surface variant. The authors show that a dual-length-scale texture—micro-rails plus a nanotextured conformal oil layer—creates a liquid surface whose apparent wettability can be patterned by the underlying solid fraction, and droplets move on it with almost no pinning: hysteresis about a degree, spontaneous uphill and inverted transport, and capture of impacting drops. That combination is new, and the visual evidence is strong.\n\nWhat is actually new: prior self-propelled droplet work on solid gradients (cactus spines, butterfly wings, chemical gradients) was limited by pinning; here the droplet never touches the solid, and the gradient lives in a liquid-on-liquid wetting energy. The force-balance model in the SI is a reasonable scaling model—Fd ~ γ_oa α R^2, Fv ~ µ_o f_s v R ln(R/L_o), giving v ∝ γ_oa α R/(µ_o f_s). The measured speed dropping as 1/f_s and rising with α and R supports the mechanism. The pinning-force measurement from critical tilt angles gives F_p ≈ 1.1 µN, order-of-magnitude consistent with the hysteresis estimate of 0.48 µN.\n\nThe model does have soft joints. First, the key wetting law (Eq. 2) comes from the authors' own prior paper and uses fl = fs + 0.059; the correction is small and plausibly explained by coating thickness, but it is a fit, so the driving force is not fully first-principles. Second, the footprint aspect ratio k in Fd = 8kγ_oa αR^2 is never reported; the force values in Fig. 3c depend on it. Third, and more substantively, the stress-test is right: SI §1.4 reports that for fs > 0.7 the air pockets under the droplet sometimes fill with oil, and the apparent contact angle drops to θe, losing its fs dependence. The transport substrate's fs goes to 1, so once a droplet's footprint moves into the filled regime, the local wettability gradient vanishes. The claim of sustained transport over the full pattern is therefore not fully backed—either the data effectively stop before the endpoint or the model overestimates the driving force in the terminal regime. That does not kill the low-to-mid-fs transport, which is the core result, but it should be re-scoped or measured directly.\n\nThis paper is for anyone working on droplet transport, microfluidics, fog harvesting, or lubricant-infused surfaces. It deserves a serious referee: the experiments are reproducible and the mechanism is worth the field's attention. My recommendation: send it out, but ask for the missing k, a direct check of whether oil fills the air pockets during transport, and a revised statement about how far along the pattern transport is actually sustained.","headline":"Liquid-on-liquid wettability gradients largely remove pinning and give real, long-range droplet transport; the qualitative mechanism holds up, but the high-fs endpoint and the fitted parameters need attention before the quantitative story is clean.","tokens_in":11033,"tokens_out":2230,"would_cite":true,"duration_ms":23649,"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 shaped liquid surface formed by nanotextured rails imbibed with silicone oil creates a liquid-on-liquid wetting state with sub-degree contact-angle hysteresis and a wettability gradient strong enough to self-propel water droplets…","keywords":["self-propelled droplet transport","shaped liquid surfaces","wettability gradient","liquid-on-liquid wetting","contact angle hysteresis","slippery liquid-infused surfaces","droplet impact capture","microfluidics"],"falsifier":"Measure the apparent contact angle of a droplet at positions along the transport gradient, especially in the $f_s > 0.7$ region, while simultaneously imaging the substrate from below. If the air pockets are replaced by oil at any position, the contact angle should drop from the Eq. 2 value to $\\theta_e = 108.4^\\circ$, and the droplet speed should deviate from $v \\propto 1/f_s$ — directly contradicting the claim of solid-free, mixed-state transport across the full gradient.","tokens_in":9980,"feed_emoji":"💧","tokens_out":4363,"duration_ms":43225,"temperature":0.7,"pith_summary":"The paper claims that replacing the solid contact underneath a droplet with a shaped liquid surface can eliminate the main obstacle to self-propelled droplet transport: contact-angle hysteresis and pinning. By coating a micro-textured solid with a thin, conformal silicone-oil layer, the droplet sits on a composite liquid-and-air surface with hysteresis below one degree. A gradual change in the underlying rail density creates a wettability gradient that drives the droplet without any external energy input. The authors show that the resulting transport is sustained, controllable in speed, works uphill, and can capture impacting droplets even on inverted surfaces. The central message is that liquid-on-liquid wetting, not solid-liquid contact, can be engineered to do the work of droplet actuation.","feed_headline":"Shaped liquid surfaces make droplets run uphill with no energy input","feed_subtitle":"A dual-scale oil coating cuts contact-angle hysteresis below 1 degree, so a wettability gradient propels drops.","key_machinery":"The load-bearing object is the shaped liquid surface: a solid micro-structure (rails, 60 µm high with spacing 75 µm) conformally coated with a hydrophobic nanoparticle layer and imbibed with 20 cSt silicone oil, so that the droplet rests on a composite surface of oil and trapped air. The apparent contact angle is set by the effective Young's law for liquid-on-liquid wetting, $\\cos\\theta_e = (\\gamma_{oa}-\\gamma_{wo})/\\gamma_{eff}$, combined with a Cassie-Baxter-type area fraction $f_l$. The mechanism that carries the argument is the force balance: the wettability gradient produces a driving force $F_d = 8k\\gamma_{oa}\\alpha R^2$ along the droplet perimeter, while dissipation at the moving contact line, $F_v \\propto f_s \\mu_o v R \\ln(R/L_o)$, provides the dominant resistance, yielding $v \\propto \\gamma_{oa} \\alpha R/(\\mu_o f_s)$. This single scaling law links the observed velocity, the measured driving and pinning forces, and the threshold gradient for motion.","core_discovery":"The paper establishes that a dual-length-scale substrate — a nanoparticulate coating that holds silicone oil, superimposed on larger micron-scale rails whose solid fraction $f_s$ varies in space — forms a shaped liquid surface on which water droplets move spontaneously. The apparent contact angle follows a liquid Cassie-Baxter relation, $\\cos\\theta = f_l \\cos\\theta_e - (1-f_l)$ with $f_l \\approx f_s + 0.059$, where $\\theta_e = 108.4^\\circ$ is the flat liquid-surface contact angle. Because the liquid surface has contact-angle hysteresis below one degree, the pinning force is tiny ($F_p \\approx 1.1 \\, \\mu\\text{N}$ measured), so even weak wettability gradients, down to $\\alpha \\approx 0.03 \\, \\text{mm}^{-1}$, overcome pinning and propel droplets over centimetre distances. Balancing the capillary driving force $F_d \\propto \\gamma_{oa} \\alpha R^2$ against viscous dissipation from the droplet contact line gives the terminal velocity $v \\propto \\gamma_{oa} \\alpha R / (\\mu_o f_s)$, matching the observed slowdown with increasing rail fraction. The same surface captures and transports impacting droplets, including hanging droplets on inverted substrates, because the liquid layer increases dissipation and normal adhesion compared with a superhydrophobic solid.","pith_inferences":["A direct corollary not spelled out in the paper is that the driving force scales with the oil-air surface tension $\\gamma_{oa}$ but not the water-oil tension, so choosing a lower-$\\gamma_{oa}$ lubricant should increase droplet speed at fixed gradient; this is a testable prediction that follows from the model.","The observed failure of air pockets at $f_s > 0.7$, where oil fills the gaps and the contact angle drops to $\\theta_e$, implies a design constraint: wettability-gradient devices should keep the operating rail fraction below this threshold if solid-free transport and the $1/f_s$ scaling are to be maintained.","The paper's force balance neglects inertial effects during impact; extending the model to include droplet deformation and oil-layer displacement during capture could predict the maximum impact velocity for which the surface still captures rather than bounces a droplet.","Because the apparent contact angle is set by the liquid surface rather than the solid chemistry, the same design could be adapted to transport other immiscible liquids, provided the lubricant has negative spreading power on the transported liquid."],"forward_implications":["If the scaling law is correct, droplet speed can be tuned continuously by adjusting the local rail fraction gradient and the lubricant viscosity, without any external actuation.","The low pinning force means that much weaker wettability gradients than those needed on superhydrophobic solids can drive transport, enabling longer travel distances per unit gradient.","Because the transport works on inverted and inclined surfaces, the design should allow droplet collection and removal in orientations where gravity opposes motion, such as fog harvesting or condensation management.","The ability to capture impacting droplets and then move them uphill suggests that shaped liquid surfaces can combine drop capture, coalescence, and directional delivery in a single passive device.","The rate $v \\propto 1/f_s$ gives a practical design rule: to slow droplets down, increase the local solid fraction, which the paper demonstrates over a range of volumes and gradients."],"supporting_citations":[{"why":"Establishes that silicone oil completely wets water and encases a water droplet, isolating it from the solid surface.","marker":"[19, 20]"},{"why":"Provides the effective Young's law for liquid-on-liquid wetting, which defines the flat-surface contact angle $\\theta_e$.","marker":"[21, 22]"},{"why":"Shows that a porous nanoparticle coating retains silicone oil and provides a continuous liquid surface, enabling the shaped liquid-surface concept.","marker":"[23, 24]"},{"why":"Documents air pockets trapped beneath droplets on such surfaces, justifying the composite liquid-and-air state used in the model.","marker":"[25]"},{"why":"Supplies the Cassie-Baxter framework that the paper adapts to a composite liquid-air surface.","marker":"[26, 27]"},{"why":"Derives the apparent contact angle relation for lubricant-impregnated surfaces, which is used as Eq. 2 with the corrected liquid fraction.","marker":"[28]"},{"why":"Provides the standard impact-bounce analysis that motivates the comparison with superhydrophobic surfaces for droplet capture.","marker":"[29, 30]"},{"why":"Identifies viscous friction in the oil layer during droplet impact, supporting the capture mechanism on liquid surfaces.","marker":"[31]"}],"fun_headline_variants":["Dual-scale oil rails let drops self-propel without power","Liquid-coated microrails send droplets uphill passively","Wettability gradient on oil surface moves drops for free","Self-driven droplets ride shaped liquid tracks"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole argument depends on the droplet remaining in a mixed liquid-and-air Cassie-Baxter state, with stable air pockets underneath it, so that the apparent contact angle follows Eq. 2 and the droplet never touches the solid; the paper itself notes that for $f_s > 0.7$ the air pockets are sometimes filled with oil, which would change the contact angle to the flat-liquid value and could weaken the gradient that drives motion.","fun_headline_variants_meta":{"raw":{"variants":["Dual-scale oil rails let drops self-propel without power","Liquid-coated microrails send droplets uphill passively","Wettability gradient on oil surface moves drops for free","Self-driven droplets ride shaped liquid tracks"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000195,"raw_usage":{"total_tokens":1399,"prompt_tokens":1030,"completion_tokens":369,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":646,"completion_tokens_details":{"reasoning_tokens":306}},"tokens_in":646,"tokens_out":369,"duration_ms":4530,"temperature":1.0,"reasoning_tokens":306,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T15:16:39.429204+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the apparent contact angle of a droplet at positions along the transport gradient, especially in the $f_s > 0.7$ region, while simultaneously imaging the substrate from below. If the air pockets are replaced by oil at any position, the contact angle should drop from the Eq. 2 value to $\\theta_e = 108.4^\\circ$, and the droplet speed should deviate from $v \\propto 1/f_s$ — directly contradicting the claim of solid-free, mixed-state transport across the full gradient.","supporting_citations":[{"cited_title":"B., Yang, S","cited_arxiv_id":null,"evidence_quote":"Documents air pockets trapped beneath droplets on such surfaces, justifying the composite liquid-and-air state used in the model."},{"cited_title":"V., Wells, G","cited_arxiv_id":null,"evidence_quote":"Derives the apparent contact angle relation for lubricant-impregnated surfaces, which is used as Eq. 2 with the corrected liquid fraction."},{"cited_title":"& Rothstein, J","cited_arxiv_id":null,"evidence_quote":"Identifies viscous friction in the oil layer during droplet impact, supporting the capture mechanism on liquid surfaces."}],"review_version":1}