Microbubble flows in superwettable fluidic channels

The control of bubble adhesion underwater is important for various applications, yet the dynamics under flow conditions are still to be unraveled. Herein, we observed the wetting dynamics of an underwater microbubble stream in superwettable channels. The flow of microbubbles was generated by integrating a microfluidic device with an electrochemical system. The microbubble motions were visualized via tracing the flow using a high-speed camera. We show that a vortex is generated in the air layer of the superaerophilic surface under laminar conditions and that the microbubbles are transported on the superaerophilic surface under turbulent conditions driven by the dynamic motion of the air film. Furthermore, microbubbles oscillated backward and forward on the superaerophobic surface under turbulent conditions. This investigation contributes to our understanding of the principles of drag reduction through wettability control and bubble flow.

Nature offers us ideas for the design of materials with superwettability. 1 In superwettable systems, the wetting of air underwater has generated interest recently. [2][3][4][5][6] For example, penguin feathers are superaerophilic, with an air layer forming on the surface underwater, which allows penguins to swim in the sea with small amounts of drag. 2,3 Inspired by this, researchers have theoretically and/or experimentally studied the inuence of wettability on drag reduction underwater. [4][5][6] In addition, sh scales are superaerophobic, which offers the idea of designing no-bubble adhesion electrodes that demonstrate high and stable oxygen evolution reaction performance. 7,8 However, despite the development of superwettable materials for the controllable adhesion of air and/or bubbles underwater, 9 the wetting dynamics of bubbles under ow conditions, which we must consider in real environments, have not been investigated.
Herein, we generated microbubble ows parallel to superwettable substrates inside a microuidic device 10,11 and studied the wetting dynamics through integrating an electrochemical setup 12 with a microuidic device, as shown in Fig. 1. The bubbles were formed through the electrolysis of water (see the ESI †). Two platinum plates were used: one as the working electrode and the other as the counter electrode. To increase the electrical conductivity, 2.0 mM K 2 SO 4 was added to the water. The bath water-vapor interfacial tension, g LV , was 71.9 AE 3.6 mN m À1 (n ¼ 15) and the pH of the water was 7.8. We applied a current of $0.25 mA cm À2 to generate microbubbles with a diameter of 463.9 AE 245.1 mm (n ¼ 120). The microuidic device was generated using a 3D printer and connected to a water-ow generator (see the ESI † for the dimensions of the device). The microbubbles generated around the electrodes moved in the direction of the water ow and the coated substrates were placed parallel to the ow.
We used the microbubbles as tracers and analyzed their ow as well as that of the water (i.e. microbubble image velocimetry), as shown in Fig. 2. We controlled the Reynolds number, Re ¼ 4Q(pDn) À1 (Q is the ow rate of the water, D is the tube diameter, and n is the kinetic viscosity of the water). Laminar ow was obtained at Re ¼ 79.21 and turbulent ow was obtained at Re ¼ 396.06 ( Fig. 2A). Under laminar ow conditions, the ow speed was nearly constant and the ow direction was close to perpendicular to the substrate (4 z 0, where 4 is the angle between the microbubble direction of movement and the width direction of the substrate) in all areas; this behavior was timeindependent ( Fig. 2B and C). Under turbulent ow conditions, the ow speed was not constant, and the ow direction was unstable (4 uctuated between À180 and 180 ) in all areas. We conrmed that the separation of ow did not occur, at least during the observation period, since the ow direction was parallel to the superwetting microuidic device.
We then prepared substrate coatings with superaerophilicity and superaerophobicity. Superaerophilic substrates were fabricated according to our previous study. 12 Concisely, a glass plate was dip-coated with a mixture of zinc oxide micro-tetrapod powder for surface roughening and polydimethylsiloxane for aerophilization. Superaerophobic surfaces were prepared through modifying a glass substrate with hydroxy groups using an aqueous potassium hydroxide solution. 13 The wettability of the superaerophobic surfaces in relation to bubbles was conrmed via measuring the underwater bubble contact angle (q); the results are shown in Fig. 3. We calculated the adhesion forces of bubbles, F adh ¼ pl 2 g LV (1 + cos q)/4, 14 where l is the bubble-solid adhesion length. On the superaerophilic surface, the adhesion force was 3.2 Â 10 3 mN, and on the superaerophobic surface the force was 4.37 mN for 6 mL bubbles. In Fig. 4, we observed air lm formation on superaerophilic surfaces under laminar and turbulent ow conditions. As we have previously shown, when microbubbles are vertically deposited on superaerophilic surfaces, a uniform air layer is formed. 10 In the present study, under both laminar and turbulent ow conditions, a uniform air layer formed on the superaerophilic surfaces, but the air layers grew non-uniformly with the deposition of microbubbles owing to Rayleigh-Taylor instability 14 (Fig. 4A and B). In all ve independent observations, the shape of the air layer was non-uniform; thus, the ow of microbubbles inuenced the shape of the air layer. However, bubbles with l ¼ 4-7 mm formed on the surfaces under both laminar and turbulent ow conditions.
Aer aging for 1000 s, a continuous air lm formed on the superaerophilic surfaces under turbulent conditions. However, the shape was unstable and changed with time (Fig. 4D). In Fig. 4C and E, we observe the formation of a vortex on the hemispherical air lm under laminar ow conditions (see Movie S1 †). This phenomenon is interesting because under laminar ow conditions a vortex should not be generated  ( Fig. 2A); this cannot be explained using Bernoulli's theorem 15 and the generation of a vortex suggests the separation of ows, which works to decrease ow resistance at the interface. Vortex generation may be due to the coalescence of microbubbles with the air layer, causing a change in the curvature of the hemispherical air lm. This, in turn, would result in a change in the Laplace pressure of 2Dkg LV , where Dk is the change in curvature. There is a uctuation in the vertical force torque to generate the vortex, and the force should be balanced by a Kutta-Joukowski force in the form of 2g LV dk/dt z rGU, where r is the density of ows, G is the vortex constant, and U is the velocity of the constant laminar ow. 16 In Fig. 4D and F, we observe that microbubbles on the air lm were transported as the shape of the air lm dynamically changed to a wave-like nature; however, the microbubbles and air lm did not coalesce (see Movie S2 †). This indicates that a thin water layer exists between the microbubbles and the air lm to prevent coalescence, whereas microbubbles are trapped on the air lm by the buoyancy force of the microbubbles, which z(Dr)Ug, where Dr is the difference in densities between a bubble and water, U is the volume of a microbubble, and g is gravitational acceleration.
We then observed the dynamics of the microbubbles on the superaerophobic surfaces (Fig. 5). As we have previously shown,  when microbubbles are vertically deposited on superaerophobic surfaces, they are uniformly deposited on the surface and have a spherical shape. 12 Under both laminar and turbulent ow conditions, microbubbles were deposited on the superaerophobic surfaces with spherical shapes but with nonuniform deposition (Fig. 5A and B). We then observed the motion of bubbles in contact with the superaerophobic surfaces. Under laminar ow conditions, microbubbles adhering to the surface moved in the direction of the ow (Fig. 5C and Movie S3 †). In contrast, turbulent ow conditions caused the microbubbles to oscillate backward and forward ( Fig. 5D and Movie S4 †). The velocimetry proles in Fig. 5E and F conrm that the bubble motion is linear in time under laminar ow, but it varies under turbulent ow (with the velocity periodically becoming negative). Despite the periodic negative velocity under turbulent ow conditions, the bubbles go forwards in the ow direction, which is not due to the laminar boundary but because the turbulent ow has more positive components than negative ones. This is because the length of positive motion under turbulent ow conditions increases with the size of the bubbles, obeying Newton's viscosity law. 17 Thus, we conrmed that the motion of bubbles on superaerophobic surfaces is inuenced by the ow conditions. The bubble motion distance on superaerophobic surfaces increased with bubble diameter.

Conclusions
We investigated the wetting dynamics of microbubbles on superwettable surfaces under laminar and turbulent ow conditions. The microbubbles were non-uniformly deposited on the surfaces and the motion of the bubbles was inuenced by the substrate wettability and the ow conditions. To discuss air adhesion on superwettable surfaces, the inuence of ow must be considered, which strongly inuences the hysteretic behavior of bubbles. 18 For instance, vortex ow formation may be a crucial factor in deciding whether air can adhere to a solid or not; Taylor instability of the air layer dynamically changes the curvature and/or air/water contact area, strongly affecting the balance of solid-liquid-air interfacial energies. Moreover, the direct in situ observation of bubble ow will help create an understanding of the interfacial phenomena demonstrated by penguins, sh, and other swimmers in nature. In addition, this work may be helpful for understanding the inuence of bubble leakage or cavitation on microuidic systems. [19][20][21]

Conflicts of interest
The authors declare no conicts of interest associated with this work.