Start-up dynamics of a reactive droplet: three-dimensional theory, contact-line stability, inertial limits, and experimental validation

A liquid droplet containing a reagent that gradually renders its solid substrate less wettable (a reactive, or autophobic, droplet) spontaneously starts to run across the substrate and eventually reaches a steady speed. The steady motion was explained by Brochard-Wyart and de Gennes (1995), and Shanahan and de Gennes (1997) analyzed the start-up: after deposition the drop first shrinks axisymmetrically as the reaction proceeds; once a direction is selected by a slight asymmetry, the center of mass accelerates at a constant rate $\gamma_1/(2\zeta\tau)$ set by the wettability-change amplitude $\gamma_1$, the contact-line friction coefficient $\zeta$, and the reaction time $\tau$, up to a crossover time $t_0\approx 2x/V_0$, after which the speed approaches the steady value $V_0=(\gamma_1 x/\zeta\tau)^{1/2}$ for a drop of half-width $x$; an attempted generalization to an axisymmetric drop gives a speed reduced by a factor $1/\sqrt{2}$. The authors themselves flagged the remaining questions, and none of them is resolved in the literature: the start-up model is a crude two-dimensional ribbon model resting on small contact angles, lubrication theory, and linear reaction kinetics; the contact-line stability argument covers only infinitesimal distortions at contact angles of order unity; inertia was bounded only for one representative parameter set; and the acceleration phase, the crossover time, and the geometry dependence of the speed have never been measured. The problem is to establish the start-up dynamics of a reactive droplet — a three-dimensional theory of the transient from deposition through shrinking, symmetry breaking, and acceleration to steady motion; contact-line stability at finite distortion amplitude; the inertial regime and its criterion; and experimental validation of the predicted constant-acceleration phase, $t_0$, and axisymmetric speed reduction.

Background

When a liquid drop is deposited on a solid it partially wets, it relaxes toward an equilibrium contact radius and contact angle $\theta_0$ set by Young's law, the force balance between the three interfacial tensions (the liquid–vapor surface tension $\gamma$, the solid–vapor tension $\gamma_S$, and the solid–liquid tension $\gamma_{SL}$) at the contact line — the curve where liquid, solid, and vapor meet. The relaxation is resisted by viscous dissipation concentrated near the contact line; in the lubrication approximation, the standard long-wavelength treatment of small-angle wetting flows reviewed by de Gennes (1985), this dissipation is represented as a friction force per unit length of contact line proportional to the line speed, with a friction coefficient $\zeta\sim\eta\ell/\theta^3$ ($\eta$ the viscosity and $\ell$ a logarithmic factor), so that a non-reactive drop relaxes exponentially on a time scale $T=\eta\ell/(\gamma\theta^3)$.

A reactive or autophobic droplet changes this picture qualitatively: the liquid carries a reagent that reacts with the substrate and renders it less wettable by that same liquid, so the quantity $\gamma_S-\gamma_{SL}$ entering Young's law decreases with contact time. Experiments by Bain, Burnett-Hall and Montgomerie (1994) — decane drops carrying a fluorinated fatty acid on glass — and by Domingues Dos Santos and Ondarçuhu (1995) — alkane drops carrying a reactive silane — showed that such drops spontaneously run across the substrate at steady speeds of order millimeters to centimeters per second, leaving a less-wettable trail behind them. Brochard-Wyart and de Gennes (1995) explained the steady motion: the advancing edge continually meets fresh substrate with a small target equilibrium angle, while the receding edge sits on reacted substrate with a larger one; the Laplace pressure (the pressure jump across the curved liquid–vapor interface, which equalizes the dynamic contact angles across the drop) then produces a net capillary driving force, balanced in the steady state by contact-line friction. For a two-dimensional drop (an elongated ribbon-like drop, of half-width $x$) with linearized reaction kinetics — the wettability change growing as $u(t)=t/\tau$, where $\tau$ is the reaction time and $\gamma_1 u(t)$ the amplitude of the surface-tension change — this gives the steady speed

$$V_0=\left(\frac{\gamma_1 x}{\zeta\tau}\right)^{1/2}.$$

Shanahan and de Gennes (1997) analyzed what happens right after deposition, before steady motion is established. First, before any direction is selected, the drop shrinks axisymmetrically: as the reaction raises the target equilibrium angle, the contact radius retreats from its initial value in three successive regimes (quadratic, then linear, then asymptotic approach to a final radius $x_\infty$ on the reaction time scale $\tau$). Second, once a slight asymmetry (substrate heterogeneity or a slight tilt) selects a direction, the two edges — exposed to the reaction for different times — experience different driving forces, and the center of mass accelerates at a constant rate

$$\frac{d^2\bar{y}}{dt^2}=\frac{\gamma_1}{2\zeta\tau},$$

that is, the speed grows linearly in time, until a crossover time

$$t_0\approx\frac{2x}{V_0}$$

— roughly when the drop has advanced by its own length and the rear edge sits on fully reacted substrate — after which the speed approaches $V_0$. Third, the authors gave an attempted generalization of the steady speed to an axisymmetric drop of final contact radius $x_\infty$ and contact angle $\theta$,

$$V_{03D}\approx\left(\frac{\gamma_1\theta x_\infty}{6\eta\ell\tau}\right)^{1/2},$$

smaller by a factor $1/\sqrt{2}$ than the two-dimensional prediction under the corresponding identification of the geometric parameters. Fourth, they estimated the ratio of inertial to driving force at $t_0$, $f_{IN}/f_D\approx\rho x^2\theta_I/(3\zeta t_0)$ ($\rho$ density, $\theta_I$ initial contact angle), finding approximately $0.2$ for representative values ($\rho=1\ \mathrm{g\,cm^{-3}}$, $\theta_I=0.5$, $\zeta=80$ cP $=0.08\ \mathrm{Pa\,s}$), so that inertia can be neglected for typical small drops. Fifth, they argued that small transverse distortions of the contact line, of wavelength $\lambda$, decay, with relative amplitude $\delta x/\lambda\sim(\gamma_1/\gamma)(T(\lambda)/\tau)\ll1$, where $T(\lambda)\sim\eta\lambda/\gamma$ is the capillary relaxation time of the distortion — so no contact-line instability is expected, but only for small distortions and contact angles of order unity.

All five results rest on idealizations the authors themselves flagged: the start-up model is two-dimensional; the axisymmetric steady speed rests on small contact angles, lubrication theory, and linear reaction kinetics; the stability argument covers only infinitesimal distortions at $\theta\sim1$; inertia is bounded only for one representative parameter set; and no experiment had — and none since has — measured the acceleration phase, the crossover time $t_0$, or the geometry dependence of the speed. The question thus remains: what are the actual start-up dynamics of a reactive droplet — the complete transient from deposition, through shrinking, symmetry breaking, and acceleration, to steady self-propelled motion — in three dimensions, with what contact-line stability, in what inertial regime, and with what experimental signatures?

A modern theoretical line has since reformulated the steady-state side of this physics: Thiele, John and Bär (2004) and John, Bär and Thiele (2005) modeled chemically driven running droplets by coupled evolution equations for the droplet profile and an adsorbate coverage field on the substrate (a thin-film gradient-dynamics model), reproducing both experimentally observed velocity regimes and analyzing the bifurcation between sitting and running drops; they also questioned the equal-dynamic-contact-angle assumption underlying the original steady-state theory. These works address the steady running state, not the start-up transient analyzed by Shanahan and de Gennes.

Problem Statement

Establish the start-up dynamics of a reactive (autophobic) droplet on a solid substrate — the complete transient from deposition, through axisymmetric shrinking and symmetry breaking, to steady self-propelled motion — and test it experimentally. The source theory (Shanahan and de Gennes, 1997) predicts, within a two-dimensional (elongated, ribbon-like drop), small-contact-angle, lubrication (long-wavelength), linear-reaction-kinetics, zero-inertia description: a constant-acceleration phase, with the center-of-mass position $\bar{y}$ obeying $d^2\bar{y}/dt^2=\gamma_1/(2\zeta\tau)$, where $\gamma_1$ is the amplitude of the reaction-induced surface-tension change, $\zeta$ the contact-line friction coefficient (the viscous friction force per unit length of contact line per unit line speed), and $\tau$ the reaction time (the wettability change growing as $u(t)=t/\tau$); a crossover time $t_0\approx2x/V_0$ — roughly the time for the drop to advance by its own length — after which the speed approaches the steady value $V_0=(\gamma_1 x/\zeta\tau)^{1/2}$ for a drop of half-width $x$; and an axisymmetric steady speed reduced by a factor $1/\sqrt{2}$ relative to the two-dimensional value under the corresponding identification of the geometric parameters. The open problem, posed by the paper's own stated limitations, has four jointly supporting parts:

(a) Three-dimensional theory. Replace the crude two-dimensional start-up model by an analysis of the full three-dimensional geometry of the advancing and receding edges, including the transition from the axisymmetric shrinking drop to a running drop, giving the transient droplet shape and center-of-mass motion and a justified steady speed for axisymmetric drops — confirming, correcting, or replacing the predicted $1/\sqrt{2}$ reduction.

(b) Contact-line stability. Decide, beyond the small-distortion, order-one-contact-angle estimate, whether transverse (contact-line-distortion) instabilities of the moving or shrinking reactive droplet exist for finite distortion amplitudes and general contact angles and reaction rates, and if so, characterize their growth.

(c) Inertial regime. Determine the dimensionless criterion — in terms of droplet size, density, viscosity, and reaction time — separating the viscous start-up analyzed in the source from regimes where inertia is significant, and derive the corresponding modified start-up dynamics.

(d) Experimental validation. Measure the start-up transient — the acceleration phase, the crossover time $t_0$, the approach to steady speed, and the speed's dependence on droplet geometry, comparing elongated (ribbon-like) and axisymmetric drops — in a system where the propulsion is demonstrably driven by the reaction-induced wettability change, and test the theoretical predictions, including whether the acceleration is constant and what sets its magnitude and duration.

These parts jointly support one objective: a self-consistent quantitative description of how a reactive droplet starts to move, its stability and inertial limits, together with, or reconciled against, experimental measurement of the predicted acceleration phase, crossover time, and geometry-dependent steady speed.

References

1. M. E. R. Shanahan and P.-G. de Gennes, 'Start-up of a reactive droplet', Comptes Rendus de l'Académie des Sciences — Series IIB — Mechanics-Physics-Chemistry-Astronomy 324(4), 261-268 (1997). DOI: 10.1016/S1251-8069(99)80032-3, https://doi.org/10.1016/S1251-8069(99)80032-3. (Primary source; publisher page paywalled.)
2. F. Brochard-Wyart and P.-G. de Gennes, 'Spontaneous motion of a reactive drop', C. R. Acad. Sci. Paris, Sér. IIb 321, 285-288 (1995). No DOI is registered for this article.
3. C. D. Bain, G. D. Burnett-Hall and R. R. Montgomerie, 'Rapid motion of liquid drops', Nature 372, 414-415 (1994). DOI: 10.1038/372414a0, https://doi.org/10.1038/372414a0
4. F. Domingues Dos Santos and T. Ondarçuhu, 'Free-Running Droplets', Physical Review Letters 75, 2972-2975 (1995). DOI: 10.1103/PhysRevLett.75.2972, https://doi.org/10.1103/PhysRevLett.75.2972
5. P.-G. de Gennes, 'Wetting: statics and dynamics', Reviews of Modern Physics 57, 827-863 (1985). DOI: 10.1103/RevModPhys.57.827, https://doi.org/10.1103/RevModPhys.57.827
6. P.-G. de Gennes, 'The dynamics of reactive wetting on solid surfaces', Physica A: Statistical Mechanics and its Applications 249, 196-205 (1998). DOI: 10.1016/S0378-4371(97)00466-4, https://doi.org/10.1016/S0378-4371(97)00466-4. (Full text paywalled.)
7. P.-G. de Gennes, 'Running droplets in a random medium', C. R. Acad. Sci. Paris, Sér. IIb 327, 147-154 (1999). DOI: 10.1016/S1287-4620(99)80024-6, https://doi.org/10.1016/S1287-4620(99)80024-6
8. M. E. R. Shanahan, 'Wetting Dynamics with Variable Interfacial Tension', Oil & Gas Science and Technology 56(1), 83-88 (2001). DOI: 10.2516/ogst:2001010, https://doi.org/10.2516/ogst:2001010.
9. U. Thiele, K. John and M. Bär, 'Dynamical model for chemically driven running droplets', Physical Review Letters 93, 027802 (2004). DOI: 10.1103/PhysRevLett.93.027802, https://doi.org/10.1103/PhysRevLett.93.027802; preprint: arXiv:cond-mat/0403043, https://arxiv.org/abs/cond-mat/0403043
10. K. John, M. Bär and U. Thiele, 'Self-propelled running droplets on solid substrates driven by chemical reactions', The European Physical Journal E 18, 183-199 (2005). DOI: 10.1140/epje/i2005-10039-1, https://doi.org/10.1140/epje/i2005-10039-1; preprint: arXiv:cond-mat/0506326, https://arxiv.org/abs/cond-mat/0506326
11. S.-W. Lee and P. E. Laibinis, 'Directed Movement of Liquids on Patterned Surfaces Using Noncovalent Molecular Adsorption', Journal of the American Chemical Society 122, 5395-5396 (2000). DOI: 10.1021/ja994076a, https://doi.org/10.1021/ja994076a.
12. S.-W. Lee, D. Y. Kwok and P. E. Laibinis, 'Chemical influences on adsorption-mediated self-propelled drop movement', Physical Review E 65, 051602 (2002). DOI: 10.1103/PhysRevE.65.051602, https://doi.org/10.1103/PhysRevE.65.051602.
13. S. Arscott, 'Dynamic Chemically Driven Dewetting, Spreading, and Self-Running of Sessile Droplets on Crystalline Silicon', Langmuir 32(48), 12611-12622 (2016). DOI: 10.1021/acs.langmuir.6b03287, https://doi.org/10.1021/acs.langmuir.6b03287.
14. Y. Sumino, N. Magome, T. Hamada and K. Yoshikawa, 'Self-Running Droplet: Emergence of Regular Motion from Nonequilibrium Noise', Physical Review Letters 94, 068301 (2005). DOI: 10.1103/PhysRevLett.94.068301, https://doi.org/10.1103/PhysRevLett.94.068301.
15. H. Taitelbaum, 'Statistical Physics Perspective on Droplet Spreading in Reactive Wetting Interfaces', Fluids 10(7), 170 (2025). DOI: 10.3390/fluids10070170, https://doi.org/10.3390/fluids10070170.
16. A. Benusiglio, N. J. Cira, A. Lai and M. Prakash, 'Two-component self-contracted droplets: long-range attraction and confinement effects' (2017). arXiv:1711.06404, https://arxiv.org/abs/1711.06404.
17. F. Voss and U. Thiele, 'Gradient dynamics approach to reactive thin-film hydrodynamics', Journal of Engineering Mathematics 149, article 2 (2024). DOI: 10.1007/s10665-024-10402-x, https://doi.org/10.1007/s10665-024-10402-x.