Electrolytic Bubble Actuators

• Electrolytic bubble actuators are devices that utilize controlled bubble nucleation and dynamic gas evolution in water electrolysis to generate mechanical motion.
• They leverage transient nanobubble regimes and alternating polarity techniques to achieve fast response times in the millisecond to microsecond range while managing gas removal effectively.
• Device architectures, including membrane actuators, combustion-driven systems, and microrobotic modules, demonstrate diverse applications such as fluid transport, precise dosing, and locomotion.

Electrolytic bubble actuators are devices that harness the dynamics of bubble nucleation, growth, coalescence, detachment, and ejection—usually generated at electrodes during water electrolysis—to create mechanical motion or perform tasks such as fluid transport, surface cleaning, or microfluidic actuation. Across the literature, their central technical problem is the coupling of strong pressure generation to reliable gas removal: classical electrochemical actuation offers force and precise metering, but conventional water-electrolysis actuators typically exhibit long response time because gas termination is slow, while newer regimes based on transient nanobubbles, alternating polarity electrolysis, controlled detachment, and even internal combustion in microchambers have shifted operation into the millisecond and microsecond domains (Uvarov et al., 2018, Sequeira et al., 2013, Uvarov et al., 2024).

1. Fundamental electrochemical and fluid-mechanical basis

At the most general level, electrolytic bubble actuation begins with gas evolution at an electrode, followed by bubble nucleation, diffusive growth, coalescence, and detachment. For water electrolysis, the electrode reactions are

$$2\mathrm{H}_2\mathrm{O} + 2e^- \rightarrow \mathrm{H}_2 \uparrow + 2\mathrm{OH}^-$$

at the cathode, and

$$2\mathrm{H}_2\mathrm{O} \rightarrow \mathrm{O}_2 \uparrow + 4\mathrm{H}^+ + 4e^-$$

at the anode. Gas production obeys Faraday’s law,

$n = \frac{I t}{z F},$

and, once a compliant boundary is present, the generated gas can be converted into displacement and work. In membrane devices, a parabolic deflection model gives the chamber-volume change as

$\Delta V = \frac{1}{2} \pi R^2 d_0.$

Bubble growth is often diffusion-controlled, with $r \propto t^{0.5}$, but electrolysis introduces continual gas generation, boundary-layer evolution, and electrode-specific wetting effects, so growth histories differ from those of passive supersaturated systems (Uvarov et al., 2018, Sequeira et al., 2013, Linde et al., 2018).

The short-time regime is qualitatively different from ordinary, diffusion-limited electrolysis. In microsystems driven on a time scale of $\sim 10\,\mu\mathrm{s}$, the Faradaic current remains effectively constant over the pulse, and extremely high current density can occur because the process is not limited by diffusion of electroactive species. Reported values include $j_F \sim 80\,\mathrm{A/cm}^2$, with local current density near the electrode periphery reaching $j_m \sim 600\,\mathrm{A/cm}^2$. The associated relative supersaturation can exceed $1000$, with $S_{H_2}^{m} \approx 2700$ and $$2\mathrm{H}_2\mathrm{O} \rightarrow \mathrm{O}_2 \uparrow + 4\mathrm{H}^+ + 4e^-$$0, making homogeneous nucleation possible in a thin diffusion layer of order $$2\mathrm{H}_2\mathrm{O} \rightarrow \mathrm{O}_2 \uparrow + 4\mathrm{H}^+ + 4e^-$$1–$$2\mathrm{H}_2\mathrm{O} \rightarrow \mathrm{O}_2 \uparrow + 4\mathrm{H}^+ + 4e^-$$2. On this short-time scale only nanobubbles can be formed; at later times these nanobubbles aggregate into microbubbles (Svetovoy et al., 2013).

Bubble evolution is not purely beneficial. Gas evolution enhances mass and heat transfer by thinning the diffusion boundary layer, but bubbles also lower the apparent electrical conductivity of bubble-containing electrolytes. The literature summarizes this tradeoff through effective-medium relations such as Maxwell’s equation, $$2\mathrm{H}_2\mathrm{O} \rightarrow \mathrm{O}_2 \uparrow + 4\mathrm{H}^+ + 4e^-$$3, and Bruggeman’s equation, $$2\mathrm{H}_2\mathrm{O} \rightarrow \mathrm{O}_2 \uparrow + 4\mathrm{H}^+ + 4e^-$$4, where $$2\mathrm{H}_2\mathrm{O} \rightarrow \mathrm{O}_2 \uparrow + 4\mathrm{H}^+ + 4e^-$$5 is the void fraction. For actuator design, this means the same bubbles that provide pressure, mixing, and convective renewal can simultaneously increase ohmic losses and reduce electrical efficiency (Sequeira et al., 2013).

2. Device architectures and actuation modes

The canonical microelectrolytic bubble actuator is a sealed chamber with integrated electrodes and a deformable membrane. A representative device uses a circular chamber of diameter $$2\mathrm{H}_2\mathrm{O} \rightarrow \mathrm{O}_2 \uparrow + 4\mathrm{H}^+ + 4e^-$$6 and height $$2\mathrm{H}_2\mathrm{O} \rightarrow \mathrm{O}_2 \uparrow + 4\mathrm{H}^+ + 4e^-$$7, fabricated on oxidized silicon with cured SU-8 side walls and two concentric Ti/Al electrodes, then sealed by a $$2\mathrm{H}_2\mathrm{O} \rightarrow \mathrm{O}_2 \uparrow + 4\mathrm{H}^+ + 4e^-$$8-thick PDMS membrane. The operating sequence is direct: a pulse train generates nanobubbles of $$2\mathrm{H}_2\mathrm{O} \rightarrow \mathrm{O}_2 \uparrow + 4\mathrm{H}^+ + 4e^-$$9 and $n = \frac{I t}{z F},$0, the nanobubble population raises chamber pressure, the membrane deflects upward, and, when the voltage pulses stop, rapid gas recombination reduces the pressure and returns the membrane toward its rest position. In this configuration, gas termination time can be as short as $n = \frac{I t}{z F},$1, cyclic operation at frequency up to $n = \frac{I t}{z F},$2 has been demonstrated, and the stroke can reach about $n = \frac{I t}{z F},$3 of the chamber volume with picoliter-range volume strokes (Uvarov et al., 2018).

A distinct architecture replaces simple gas-pressure loading with combustion-driven pressure release. In a chamber with a volume of $n = \frac{I t}{z F},$4 nanoliters, alternating polarity voltage pulses prepare a cloud of nanobubbles containing a stoichiometric $n = \frac{I t}{z F},$5 mixture. This cloud merges to form a microbubble that explodes, increasing the volume $n = \frac{I t}{z F},$6 times in $n = \frac{I t}{z F},$7. Reported performance includes an instantaneous force up to $n = \frac{I t}{z F},$8, the ability to move a body $n = \frac{I t}{z F},$9 times more massive than the actuator itself, and a natural response time of about $\Delta V = \frac{1}{2} \pi R^2 d_0.$0, defined by the incubation time needed to produce an exploding bubble. Reliable cyclic actuation at $\Delta V = \frac{1}{2} \pi R^2 d_0.$1 was demonstrated, limited by electrolyte aging, and no significant wear in the chamber was observed after $\Delta V = \frac{1}{2} \pi R^2 d_0.$2 explosions (Uvarov et al., 2024).

Electrolytic bubble actuation also appears in locomotion-oriented systems. In modular microrobots, planar Pt/Ti and Ni bubble-generating electrodes are patterned on cube faces. Electrolysis generates microbubbles of diameter $\Delta V = \frac{1}{2} \pi R^2 d_0.$3–$\Delta V = \frac{1}{2} \pi R^2 d_0.$4, a growing internal gas pocket raises pressure according to $\Delta V = \frac{1}{2} \pi R^2 d_0.$5, the actuated face deforms and tilts the cube, and bubble escape under the lifted edge produces stepwise propulsion. Reported values include displacement per cycle of $\Delta V = \frac{1}{2} \pi R^2 d_0.$6, an actuation rate of $\Delta V = \frac{1}{2} \pi R^2 d_0.$7, actuation force up to $\Delta V = \frac{1}{2} \pi R^2 d_0.$8 per face, and locomotion speed of $\Delta V = \frac{1}{2} \pi R^2 d_0.$9–$r \propto t^{0.5}$0 on wet surfaces (Bandari et al., 24 Aug 2025).

Architecture	Operating principle	Representative characteristics
Membrane microactuator	Alternating polarity water electrolysis	$r \propto t^{0.5}$1 response, up to $r \propto t^{0.5}$2, picoliter-range strokes
Combustion membrane actuator	Internal combustion of $r \propto t^{0.5}$3 and $r \propto t^{0.5}$4 in a microchamber	Volume $r \propto t^{0.5}$5 times in $r \propto t^{0.5}$6, force up to $r \propto t^{0.5}$7
Bubble-generating microrobot actuator	Face-selective electrolysis and bubble escape	$r \propto t^{0.5}$8 per cycle, $r \propto t^{0.5}$9, $\sim 10\,\mu\mathrm{s}$0–$\sim 10\,\mu\mathrm{s}$1

3. Nanobubbles, alternating polarity, and rapid recovery

The decisive advance in fast electrolytic actuation is the replacement of persistent microbubbles by transient nanobubbles. Alternating polarity water electrolysis with short, high-frequency voltage pulses—typically $\sim 10\,\mu\mathrm{s}$2 pulses at about $\sim 10\,\mu\mathrm{s}$3—produces only nanobubbles. Because hydrogen and oxygen are created in close proximity and confined at very high surface-to-volume ratio, the gas in nanobubbles can be terminated fast due to surface assisted reaction between hydrogen and oxygen that happens at room temperature. The reported surface-to-volume ratio is on the order of $\sim 10\,\mu\mathrm{s}$4, and gas termination occurs in less than $\sim 10\,\mu\mathrm{s}$5 when the electrical pulses stop. This is the key reason why alternating-polarity membrane actuators can operate at hundreds of hertz rather than on minute-long recovery times (Uvarov et al., 2018).

Short-time electrolysis further revealed that stoichiometric nanobubbles below a critical diameter of $\sim 10\,\mu\mathrm{s}$6 can undergo spontaneous reaction. In alternating-polarity operation at frequencies above about $\sim 10\,\mu\mathrm{s}$7, a nanobubble can collect a stoichiometric mixture before it grows large; spontaneous reaction between the gases is then observed, and such bubbles disintegrate violently, affecting the electrode surface. At lower frequencies, bubbles grow past the critical size before becoming stoichiometric and the mixture persists as visible microbubbles. This establishes a size- and frequency-dependent threshold between benign gas accumulation and energetic collapse (Svetovoy et al., 2013).

The nanobubble literature also supplies a current-threshold view of detachment. Molecular dynamics and generalized nanobubble stability theory show that there is a threshold current below which electrolytic nanobubbles reach a stable equilibrium size and contact angle, and above which they grow without bound and detach. The minimum current for bubble detachment is given analytically by

$\sim 10\,\mu\mathrm{s}$8

In the detachment regime, the radius first increases as $\sim 10\,\mu\mathrm{s}$9 and then as $j_F \sim 80\,\mathrm{A/cm}^2$0 up to bubble detachment. For actuator engineering, this provides a quantitative criterion for switching between stable pinned nanobubbles and continuous gas-release operation (Zhang et al., 2023).

A different nanoscale regime is bubble immobilization by nanoelectrolysis. Using a Pt nanoelectrode with an apex curvature radius from $j_F \sim 80\,\mathrm{A/cm}^2$1 to $j_F \sim 80\,\mathrm{A/cm}^2$2 and an alternating electric potential, a bubble produced by water electrolysis can be immobilized in the liquid when the frequency is rapidly increased to at least $j_F \sim 80\,\mathrm{A/cm}^2$3–$j_F \sim 80\,\mathrm{A/cm}^2$4. The explanatory model is an open bubble exchanging $j_F \sim 80\,\mathrm{A/cm}^2$5 and $j_F \sim 80\,\mathrm{A/cm}^2$6 with a non-uniform dissolved-gas field generated at the apex. The phenomenon does not correspond to a passive force balance against buoyancy; rather, the bubble is actively stabilized by a localized chemical environment (Hammadi et al., 2015).

4. Detachment, coalescence, and interfacial transport

Detachment is governed not only by buoyancy and capillarity but by the dynamics of the three-phase contact line. Recent experiments distinguish pinned bubbles, whose contact line remains fixed during growth, from spreading bubbles, whose contact line expands before pinning. The classical Fritz formula for the departure radius in the spreading case,

$j_F \sim 80\,\mathrm{A/cm}^2$7

underpredicts the detachment radius by a large margin when contact angle hysteresis is present. The observed mechanism is a two-stage process: spreading at the receding contact angle, then pinning, then detachment when the contact angle reaches the advancing contact angle. At $j_F \sim 80\,\mathrm{A/cm}^2$8 concentrations greater than or equal to $j_F \sim 80\,\mathrm{A/cm}^2$9, spreading bubbles become more prevalent; at concentrations less than or equal to $j_m \sim 600\,\mathrm{A/cm}^2$0, pinned bubbles dominate. For actuation, this means that bubble lifetime and departure size can be tuned by surface wetting, roughness, and electrolyte composition rather than by buoyancy alone (Demirkır et al., 2024).

Bubble–bubble interaction adds another layer of control. Under cyclic modulation of the electric potential at a Pt microelectrode in acidic electrolyte, three interaction scenarios were identified for successive $j_m \sim 600\,\mathrm{A/cm}^2$1 bubbles. The most prominent is motion reversal and coalescence: the first detached bubble initially moves upward, then reverses direction and returns to the electrode before merging with the second bubble. Schlieren visualization and modeling attribute this to the competition between buoyancy and thermocapillary effects; when the thermal boundary layer from the second bubble overlaps the underside of the first, Marangoni force can drive the first bubble downward against buoyancy. The corresponding migration velocity is expressed as

$j_m \sim 600\,\mathrm{A/cm}^2$2

This shows that detached bubbles are not necessarily irretrievable; they can be returned, merged, or repelled by spatiotemporal control of bubble generation (Bashkatov et al., 2022).

Coalescence can also alter bubble composition. During hydrogen evolution, coalescence with microbubbles can drive electrolyte droplets into the gas phase through fragmentation of a Worthington jet. The resulting droplets are typically $j_m \sim 600\,\mathrm{A/cm}^2$3, increasing to $j_m \sim 600\,\mathrm{A/cm}^2$4 just before bubble departure. Most droplets move at velocities up to $j_m \sim 600\,\mathrm{A/cm}^2$5, while rare high-speed droplets reach $j_m \sim 600\,\mathrm{A/cm}^2$6. In electrode-attached bubbles, the sprayed droplets form electrolyte puddles at the bubble–electrode contact area, affecting the dynamics near the three-phase contact line and favoring bubble detachment from the electrode. This corrects the common simplification that an electrolytically generated $j_m \sim 600\,\mathrm{A/cm}^2$7 bubble contains only hydrogen gas and vapor (Bashkatov et al., 2024).

A related controversy concerns the role of Marangoni stresses in coalescence. Recent measurements of thin-film drainage show that film morphology governs drainage more strongly than interfacial boundary conditions. Coalescence proceeds through three regimes: a visco-capillary stage with $j_m \sim 600\,\mathrm{A/cm}^2$8, an exponential decrease induced by rim stabilisation, and a final exponential relaxation governed by disjoining pressure. After rescaling with characteristic film thickness and time scale, data for different electrolytes collapse onto a single curve, indicating that electrolyte effects act only to renormalize timescales rather than alter the underlying dynamics (Palliyalil et al., 26 Jun 2026).

5. Measurement, sensing, and feedback

The field has depended on unusually diverse diagnostics because bubble actuators couple electrochemistry, elasticity, wetting, and optical interfaces. Membrane microactuators have been characterized by an interferometer and a fast camera, enabling direct measurement of membrane deflection and rapid events such as controlled explosions that push the membrane up to $j_m \sim 600\,\mathrm{A/cm}^2$9. High-speed shadowgraphy, particle tracking velocimetry, and Toepler’s schlieren technique have been used to map contactless motion reversal, flow, and temperature gradients around interacting $1000$0 bubbles (Uvarov et al., 2018, Bashkatov et al., 2022).

Direct observation of detachment physics has been improved by transparent electrodes. Transparent Pt or Ni electrodes permit imaging of the bubble contact line and inference of contact angle and volume through solutions of the Young–Laplace equation. In another context, X-ray micro-CT has validated static equilibrium models for bubbles sustained underneath perforated electrodes immersed in liquid, showing that electric field acts as an additional pressure term on the meniscus. In silicone oil, an applied voltage from $1000$1 to $1000$2 produced a protrusion shift predicted to be $1000$3 and observed as $1000$4 (Demirkır et al., 2024, Tesi et al., 2021).

Electrical sensing has made single-bubble monitoring possible without optical access. Zero-dimensional ion-sensitive field-effect transistors buried under a microbath can detect the emission of individual $1000$5 bubbles as sharp spikes in output current. Bubble size can then be inferred from electrolysis current and emission frequency using

$1000$6

In the reported setup, the estimated bubble radius was $1000$7, and bubble size dispersion was about $1000$8 for single nucleation, compared with about $1000$9 in systems with multiple nucleation centers. This suggests a practical route toward feedback-controlled actuation with electrical bubble counting rather than purely optical metrology (Clement et al., 2013).

6. Applications, integration, and design constraints

The most direct applications are in microfluidics and dosing. The millisecond membrane actuator was explicitly positioned as an advance for drug delivery and microfluidics because it combines a chamber volume of $S_{H_2}^{m} \approx 2700$0, volume strokes in the picoliter range, and highly repeatable cyclic operation under loads such as a micro-mirror imposing about $S_{H_2}^{m} \approx 2700$1 back pressure. More broadly, electric-field-mediated bubble generation and manipulation in confined geometries have been linked to on-chip valving, droplet generation, chemical delivery, digital microfluidics, emulsification, and foam generation (Uvarov et al., 2018, Chakraborty et al., 2013).

Bubble actuation also appears in interface-shaping systems not primarily designed as pumps. In bubble-assisted Liquid Hole Multipliers for noble-liquid radiation detectors, the bubble meniscus under a perforated electrode controls electron transfer and electroluminescence yield. A validated electro-hydrostatic model shows that the electric field contributes an additional pressure term,

$S_{H_2}^{m} \approx 2700$2

so increasing voltage increases meniscus protrusion and curvature. This is not an actuator in the classical membrane-displacement sense, but it is an electro-mechanical bubble interface whose equilibrium can be programmed by field and material choice (Tesi et al., 2021).

At the system level, integration has advanced to autonomous microrobots. Smartlets with all dimensions of $S_{H_2}^{m} \approx 2700$3 and below combine electrolytic bubble actuators on multiple faces with on-board energy harvesting, sensors, and a CMOS microcontroller. Actuation pulses can be selected by program or by photodetector input, allowing 2D locomotion on wet surfaces and selective docking with other modules. This demonstrates that electrolytic bubble actuators can function not only as isolated components but as embedded propulsion units in heterogeneous microsystems (Bandari et al., 24 Aug 2025).

Several design constraints recur across the literature. Current prototypes can require $S_{H_2}^{m} \approx 2700$4–$S_{H_2}^{m} \approx 2700$5 and can consume about $S_{H_2}^{m} \approx 2700$6 in high-frequency operation; the combustion actuator consumed $S_{H_2}^{m} \approx 2700$7 per explosion and $S_{H_2}^{m} \approx 2700$8 at $S_{H_2}^{m} \approx 2700$9, with the demonstration frequency restricted by electrolyte aging rather than by the chamber mechanics (Uvarov et al., 2018, Uvarov et al., 2024). Electrode architecture can be equally limiting: in 3D-printed microlattice HER electrodes, increasing electrochemically active surface area did not guarantee better performance because severe bubble trapping in complex pores offset the benefit of larger area. Across the larger bubble size range of about $$2\mathrm{H}_2\mathrm{O} \rightarrow \mathrm{O}_2 \uparrow + 4\mathrm{H}^+ + 4e^-$$00–$$2\mathrm{H}_2\mathrm{O} \rightarrow \mathrm{O}_2 \uparrow + 4\mathrm{H}^+ + 4e^-$$01, only $$2\mathrm{H}_2\mathrm{O} \rightarrow \mathrm{O}_2 \uparrow + 4\mathrm{H}^+ + 4e^-$$02–$$2\mathrm{H}_2\mathrm{O} \rightarrow \mathrm{O}_2 \uparrow + 4\mathrm{H}^+ + 4e^-$$03 of bubbles escaped through even the most open microlattices, and up to $$2\mathrm{H}_2\mathrm{O} \rightarrow \mathrm{O}_2 \uparrow + 4\mathrm{H}^+ + 4e^-$$04 of large bubbles could be trapped depending on geometry and bubble size. This suggests that efficient bubble actuators must optimize for bubble escape, not merely for gas generation or area maximization (Ferguson et al., 17 Dec 2025).

Taken together, the literature defines electrolytic bubble actuation as a family of electrochemical transduction strategies in which interfacial physics is as important as electrolysis itself. Fast operation depends on suppressing persistent microbubbles, exploiting nanobubble-mediated recombination or combustion, and controlling contact line dynamics, coalescence, and bubble traffic. Strong operation depends on converting gas generation into membrane loading, tilt, or confined pressure spikes without incurring prohibitive conductivity loss, trapping, or material degradation. The persistent open questions—especially around detachment variability, the relative roles of Marangoni stress and film morphology, and architecture-dependent gas management—remain central because they determine whether bubble generation behaves as an actuator, a parasitic loss, or both (Demirkır et al., 2024, Palliyalil et al., 26 Jun 2026).