Tackling the Short-Lived Marangoni Motion Using a Supramolecular Strategy

Introduction

The Marangoni effect, reminiscent of the “tears of wine” phenomenon, refers to the interfacial transport driven by the surface tension gradient formed between two liquids with different surface tensions.¹ The Marangoni effect has been widely exploited by living organisms, such as aquatic beetles, to escape from predators^2,3 and by artificial devices to enable an intriguing self-propulsion mechanism in which spontaneous motion results from the energy release from the physicochemical process.^4,5 However, such self-propulsion enabled by the Marangoni effect exhibits a typical characteristic of “one-shot” motion imposed by the complexity of forming and maintaining a stable local surface tension gradient.^5,6 For example, “soap boats,” the most well-known type of Marangoni motion in our daily lives,⁷ are limited by a very short lifetime ranging from seconds to minutes.¹ Although a soap boat can be propelled by the introduction of amphiphilic surfactants consisting of a hydrophilic head and a hydrophobic tail, these propellants can rapidly occupy the limited water surface and inactivate the solid/liquid interface. As a result, the interfacial surface tension gradient is lost.^1,4 To overcome the short lifetime inherent in the Marangoni motion, different strategies have been proposed. The first approach includes the optimization of the experimental parameters, such as slowing the release of surfactant,^8,9 reducing the motor size,^10,11 and increasing the water surface area, whereas the second approach depends on the use of special surfactants, such as ethanol,^4,12 acetonitrile,¹³ dimethylformamide,^13,14 camphor,^15,16 and ionic liquids.⁸ Despite extensive progress, a facile approach to overcome the one-shot challenge inherent in classical Marangoni motion is lacking.

We propose a supramolecular strategy based on the concept of supra-amphiphiles^17,18 by utilizing the intricate competition between interfacial adsorption, which is dominated by the kinetics of the involved physicochemical processes, and the simultaneous desorption of surfactant molecules, which serves to recover the surface tension gradient. To validate our hypothesis, we establish a model system resembling classical soap boat motion using an amphiphilic surfactant, sodium dodecyl sulfate (SDS), as a propellant. Furthermore, to continuously remove SDS from the interface, we take advantage of the supramolecular recognition between β-cyclodextrin (CD) and SDS to form the water-soluble CD–SDS supramolecular complex as a competitive equilibrium.^19–21 The balance of the physicochemical equilibrium between SDS adsorption and CD–SDS formation transforms the interfacial distribution of SDS at the molecular level, thereby reactivating a stable surface tension gradient and prolonging the Marangoni motion from the original 55 to 2400 s. We further demonstrate that this prolonged Marangoni motion can be harnessed to realize electricity generation.^22,23 We envision that this supramolecular strategy may advance further applications of both Marangoni motions and the concept of supra-amphiphiles.

Results and Discussion

Classical Marangoni motion

To validate our hypothesis, we used a soap boat as a model system. Generally, the soap boat was constructed from low-density material and functionalized with dish soap at one end.^1,7 Although this design can trigger random motion on the water surface, it is vulnerable to several challenges, such as edge effects caused by collision, uncertainty of trajectory, and complex fluidic impacts on velocity changes. To overcome these problems, we designed a new kind of soap boat. First, the boat hull was designed as a three-wing rotator device (Figure 1a). To support the device and serve as a rotation axis, a through hole was opened in the middle of the wing with the insertion of a free-standing copper wire. Thus, the motion trajectory of the boat under varying experimental conditions was regulated and recorded (Figure 1b). Second, a hydrogel loaded with SDS was used as the soap engine, thereby simplifying the chemical composition of the soap and the subsequent release/adsorption processes. Third, to further achieve drag reduction, the wing surface was decorated with a superhydrophobic coating (Figure 1c), which was obtained by the deposition of rough silver aggregates with hierarchical structures²⁴ and subsequent surface modification using 1-dodecanethiol (Figure 1d). The surfactant SDS was mixed within bulk acrylamide hydrogels and then loaded at the wing end of the rotator device to serve as fuel for the Marangoni motion (for fabrication details, see ).

• Download figure
• Download PowerPoint

On the surface of pure water, the newly designed soap boat exhibits the typical one-shot feature of Marangoni motion.^1,5 The velocity profile displays a dramatic change from the initial 75 to ∼30 mm/s within 5 s (details of velocity calculation are described in ), and a complete cessation of motion occurs at 55 s (Figure 2a; ). To explain the remarkable decrease in velocity, we measured the time-dependent variation in the surface tension gradient (see and for details on apparatus for the measurement of surface tension), the major driving force of the boat motion. At 5 s, the surface tension is measured to be 50 mN/m (Figure 2b), corresponding to a surface tension gradient of 16 mN/m, which is much smaller than that in the initial stage (38 mN/m). Eventually, the surface tension levels off to 34 mN/m, indicating the total loss of the surface tension gradient.

• Download figure
• Download PowerPoint

The time-dependent variation in the surface tension can be elucidated by considering the transport dynamics. As schematically illustrated in Figure 2c, initially, the release of the SDS molecules creates a surface tension valley sufficient to initiate Marangoni flow. However, the continuous adsorption of SDS at the air/water interfaces also inactivates the surface tension gradient. Concurrent with the interfacial adsorption, SDS molecules are also associated with the process of dissolving into the bulk water phase,²⁵ which is competitive with the adsorption process in depleting SDS at the interface. To probe the influence of the dissolution process on the variation in surface tension, we conducted a control experiment in which the water surface was physically separated with a plastic barrier plate into two regions (surfaces A and B), whereas the bulk solution was still connected (Figure 2d). In the first 100 s, we placed an SDS-added hydrogel onto surface A, which is associated with only a dramatic drop in surface tension. Notably, upon removing the SDS hydrogel, we observed a gradual recovery in surface tension, which is ascribed to the partial transfer of SDS from surface A to the bulk solution. On the other hand, contrary to our perception, surface B also displays a gradual decline in surface tension (Figure 2e), suggesting the steady adsorption of SDS molecules on its surface. As surface B is not in direct contact with SDS, continuous SDS adsorption should result from the underlying dynamic transportation from surface A to bulk solution. The transportation process took up to 1600 s to reach the plateau of the surface tension level, indicating that the SDS dissolution process is much less competitive in the kinetics, as the SDS adsorption was completed within 50 s (Figure 2b). We have thus demonstrated the competition between the SDS adsorption and dissolution processes, which could slowly tailor the transport dynamics of SDS in classical Marangoni motion on pure water. Therefore, to accelerate the transport dynamics and prolong Marangoni motion, we sought to enhance the SDS dissolution process in the solution phase.

Supramolecular strategy to tailor interfacial equilibrium

We next propose a facile supramolecular strategy to enhance the SDS dissolution at the interface to reactivate the desired surface tension gradient. Supramolecular recognition, such as host/guest interactions,²⁶ is an ideal molecular interference strategy to deplete SDS in bulk solution. Hence, we choose CD–SDS molecular recognition^19–21 (Figure 3a) for two reasons. First, the host molecule of CD and the guest molecule of SDS can form a water-soluble supramolecular complex with a high binding constant,^27,28 and the resulted surface tension is almost comparable with that of pure water.¹⁹ Second, because of CD–SDS molecular recognition, SDS molecules can easily be dissolved in the solution phase, thereby leading to fast recovery of the surface tension on the water surface. The presence of CD has been demonstrated to be effective in increasing the critical micelle concentration of SDS.^29,30 Overall, we can establish a favorable steady surface tension gradient by balancing the kinetics of SDS adsorption and CD–SDS recognition.

• Download figure
• Download PowerPoint

We further show that the CD–SDS recognition kinetics can be adjusted by varying the CD concentration in the solution phase (see ). , plot the variation in the surface tension and motion lifetime as a function of the CD concentration. Clearly, the maximum motion lifetime is achieved at C_CD = 10 mM, which results from the interplay between the SDS adsorption kinetics (kinetic constant, k₁) and CD–SDS binding kinetics (k₂). Initially, upon coming into contact with water, the dry hydrogel readily released SDS, and thus, the interfacial interaction is mainly dominated by the SDS adsorption process (k₁ > k₂). Indeed, the motion velocity displays a high initial value followed by a fast decrease (Figure 3b; ), which is further confirmed by the surface tension measurement (Figure 3c). As the SDS release slows down (k₁ decreases), the CD–SDS binding process becomes active (k₂ > k₁), and hence, SDS molecules can be removed more effectively (Figure 3d). This is inconsistent with the surface tension measurement, which shows a steady high level of 67 mN/m throughout the motion. Therefore, the supramolecular strategy has provided an efficient method to tailor the interfacial adsorption of surfactant molecules based on competing physicochemical equilibriums.

Visualization of Marangoni motion trajectory

The effects of supramolecular recognition on the transformation of the interfacial interaction can be visualized with thymol blue dye particles as tracers.³¹ We distributed a thin layer of thymol blue particles to fully occupy the surface of pure water, the disturbance of which could be traced by the relative motion of the particles. When the device was placed onto the water surface (; Figure 3e), the dye particles around each hydrogel were pushed away, leaving clear round water zones similar to three “footprints” at each wing end. Within 1 s, the three footprints emerged and expanded throughout almost the whole water surface. At 5 s, the dye particles were completely pushed toward the edges as an irregular thin rim. The instant diffusion of the dye particles indicates that the released SDS molecules had soon occupied the water surface. However, the situation was different for the rotation on the CD solution (C_CD = 10 mM) (; Figure 3f). The “footprints” expanded only slightly to the hydrogel side and formed a weak vortex within a circle covered by the rotator device. The dye particles were pushed away locally where the wing traveled but soon gathered near the device. Therefore, at the detection point for in situ surface tension measurement, a steady high surface tension level can be maintained (Figure 3c). Thus, the SDS release area always has a low surface tension, whereas the far water area keeps a high surface tension, thereby achieving a relatively stable surface tension gradient and continuous rotation.

Electricity generation through prolonged Marangoni motion

The prolonged Marangoni motion with a regular trajectory can be directly exploited for energy harvesting.^22,23 We applied Faraday’s law of induction to the motion system with the setup shown in Figure 4a. Three external magnets (0.3 T each) were adhered to the end of the three wings, and a solenoid coil was fixed with 3,000 turns of wire above the device at a distance of 10 mm. The generated voltage or current was detected through an electrochemical workstation. The continuous rotation of three magnets on the device wings naturally induced a time-dependent variation in the magnetic flux and electricity in the closed circuit, as evidenced by the potential-time curves shown in Figure 4b. The maximum voltage was measured to be ∼0.7 mV and gradually decayed to 0.2 mV in the long-motion phase. For each circle in which the device rotated, there are three pairs of positive/negative peaks due to the movement of the three magnets approaching and moving away from the coil. The cyclical trend in the voltage matches well with that of the velocity because the motion speed primarily determines the induced electricity intensity. In addition, extra resistance or drag was generated to decelerate the rotation when the magnets approached or moved away from the coil according to Lenz’s law. Using this method, we can harvest electrical energy through the Marangoni-effect-driven motion. By contrast, for the pure water system, the induced voltage decayed rapidly within 20 s, with a corresponding sharp drop in the voltage values (Figure 4c).

• Download figure
• Download PowerPoint

Conclusion

In summary, we have established a supramolecular strategy to overcome the limitation of the short lifetime of the Marangoni motion by breaking the physical limit of interfacial inactivation in motion with amphiphilic surfactants as a propellant. A dynamic equilibrium process is created by the CD–SDS supramolecular recognition and the SDS adsorption process. We achieved a relatively steady surface tension gradient by tailoring the kinetics of the above two competitive interfacial processes and prolonged the motion lifetime from 55 s on water to 2400 s on CD solution. The prolonged Marangoni self-propulsion is further applied in electricity generation with regulated trajectories in the presence of a magnetic field. Moreover, the prolonged Marangoni motion holds promise for providing sustainable driving forces for macroscopic building blocks for further self-assembly.

Supporting Information

Supporting information is available.

Conflicts of Interest

The authors declare no competing financial interests.