In-air fast response and high speed jumping and rolling of a light-driven hydrogel actuator

Stimuli-responsive hydrogel actuators have promising applications in various fields. However, the typical hydrogel actuation relies on the swelling and de-swelling process caused by osmotic-pressure changes, which is slow and normally requires the presence of water environment. Herein, we report a light-powered in-air hydrogel actuator with remarkable performances, including ultrafast motion speed (up to 1.6 m/s), rapid response (as fast as 800 ms) and high jumping height (~15 cm). The hydrogel is operated based on a fundamentally different mechanism that harnesses the synergetic interactions between the binary constituent parts, i.e. the elasticity of the poly(sodium acrylate) hydrogel, and the bubble caused by the photothermal effect of the embedded magnetic iron oxide nanoparticles. The current hydrogel actuator exhibits controlled motion velocity and direction, making it promising for a wide range of mobile robotics, soft robotics, sensors, controlled drug delivery and other miniature device applications.

The work presents the development and characterization of new light-driven hydrogel actuator, based on rapid phase transition (liquid to gas phase), able to jump and roll at a high speed and in air. The transition phase is generated by photothermal heating provided by a NIR laser mediated by iron oxide nanoparticles embedded in a hydrogel matrix. The authors show different motions and modality of control of the demonstrators. Respect to the state of the art of the field, the actuation here reported results impressively fast and reactive. The work has been clearly introduced and the level of innovation is clear and substantial. Finally the videos are very impressive and informative. The paper is well written, clear, and easy to follow. The language is appropriate and the paper is technologically sounding. I have just a few comments that could improve the manuscript before the publication, as follows: -Although in the methods it is explicitly reported that the "Unless otherwise mentioned, the light intensities to actuate the jumping and rolling motion of the hydrogel actuator were 1.2 and 0.3 W/cm2, respectively", I think that this should be stated in the main text when the actuation is discussed, and in all the caption of the figures in which the motion is presented.
-The expression of laser light intensity in unit of [W/cm^2] is not fully appropriate because, since the laser beam comes out from an optical fiber, the power strongly depends from the focus spot area. By the way this important parameter is not reported in the text. Probably it is not ease to define the area of IR spot, since the focalization seems to be made manually, but a range of spots areas should be provided. I suggest to use the power of the laser (expressed directly in W), which is the parameter that is controlled by the lased driver -and that can be more simply calibrated by standard procedures-and to calculate the typical power density (or light intensity) for selected applications, as reference for comparing the work with others.
-Other minor comment: in Figure 1, panel "j" and "m" should report "jumping" and "rolling" directly in the graph, for a better readability. "Among others, inspired by the motion of plants whose winding and curving are driven by the absorption or dehydration of water in the cells and tissues, …..." insert reference(s) here.
Reviewer #2 (Remarks to the Author): Li et al presented a new mechanism for hydrogel actuation and achieved ultra-high speed jumping and rolling by applying localized high-density NIR light. This is a very interesting work, overcoming the traditional drawbacks of hydrogel actuators. I recommend its publication in NC. In addition, the following comments and concerns could be addressed to improve the paper. 1 We would like to thank the Reviewers for their invaluable feedback and comments. We have addressed all comments outlined by the reviewers as detailed below point by point by significantly improving our paper. We have marked the reviewers' comments as red italic font and our responses as normal black font. The revised portion of the paper is highlighted in yellow in both the responses below and in the revised manuscript.

Reviewer 1:
The work presents the development and characterization of new light-driven hydrogel actuator, based on rapid phase transition (liquid to gas phase), able to jump and roll at a high speed and in air. The transition phase is generated by photothermal heating provided by a NIR laser mediated by iron oxide nanoparticles embedded in a hydrogel matrix. The authors show different motions and modality of control of the demonstrators. Respect to the state of the art of the field, the actuation here reported results impressively fast and reactive. The work has been clearly introduced and the level of innovation is clear and substantial. Finally, the videos are very impressive and informative.
The paper is well written, clear, and easy to follow. The language is appropriate and the paper is technologically sounding.
I have just a few comments that could improve the manuscript before the publication, as follows: 2 1. 'Although in the methods it is explicitly reported that the "Unless otherwise mentioned, the light intensities to actuate the jumping and rolling motion of the hydrogel actuator were 1.2 and 0.3 W/cm 2 , respectively", I think that this should be stated in the main text when the actuation is discussed, and in all the caption of the figures in which the motion is presented.' Response: Thank you for your kind comment. We have revised the manuscript and provided the laser power in the main text and in all caption of the figures when the actuation is discussed according to your comment. i) The NIR laser beam is divergent and the spot area is dependent on the distance from the laser to the object. Because of the invisibility of the NIR light, we estimate the laser spot area In the current study, the distance between the NIR laser and the hydrogel actuator is about 2 mm. The spot area is thus approximately 0.28 mm 2 .

'The expression of laser light intensity in unit of
ii) In addition, we have revised the manuscript and used the power of the laser (W) instead of light density (W/cm 2 ). 3 We have thus added the new Supplementary Figure 29 and the corresponding descriptions are added on page 15, line 5 of the second paragraph: "The NIR laser beam without a collimator is divergent so that the spot area on the hydrogel actuator is dependent on the distance from the laser to the object and the temperature of the hydrogel actuator is only higher than 100 °C when the distance between the NIR light and the actuator is less than 4 mm (with a spot area of approximately 0.28 mm 2 at 2 mm separation) as shown in  Figure 1, panel "j" and "m" should report "jumping" and "rolling" directly in the graph, for a better readability.' 4 Response: Thank you for your comments. For clarity, we added "Jumping" and "Rolling" to

Reviewer 2:
Li et al presented a new mechanism for hydrogel actuation and achieved ultra-high speed jumping and rolling by applying localized high-density NIR light. This is a very interesting work, overcoming the traditional drawbacks of hydrogel actuators. I recommend its publication in NC. In addition, the following comments and concerns could be addressed to improve the paper. i) This method is not limited to the small-sized beads. 5 ii) The driving force originates from the shape deformation caused by the bubble formation due to the photothermal effect of Fe3O4 nanoparticles, which not only scales with the area but also scales with the contact pressure, as shown in the following equation

'Light controlled actuation of robots has the advantage of the remote control. But in the experiments showed by the authors, the light source needs to be very close to the beads to achieve efficient actuation. What is the difference between this light-assisted technology and manual operation of the beads?'
Response: Thank you for your comments.
i) The actuation of the hydrogel actuator relies on the bubble formation caused by the photothermal effect of Fe3O4 nanoparticle and the local temperature of the hydrogel actuator under NIR irradiation has to exceed 100 °C . However, the NIR light beam utilized in the current study is divergent. We have monitored the temperatures of the hydrogel actuator under NIR illumination when changing the distances between the NIR light and the actuator. 9 As can be seen from new Supplementary Figure 29(a)-(c), the temperature of the hydrogel actuator is higher than 100 °C when the distance between the NIR light and the actuator is less than 4 mm. In order to increase the actuation distance, there are a few possible ways, which include the utilization of NIR laser with higher power, narrowing the divergent NIR light or their combination.
Among others, we have examined the possibility to increase the actuation distance by narrowing the divergent NIR beam utilized in the current study on the hydrogel actuator through the application of a collimator at the end of the NIR optical fiber, as shown in new

Supplementary Figure 28.
We also monitor the temperature of the hydrogel actuator under the irradiation of the narrowed NIR light. As can be seen in Supplementary Figure 30, once the divergent NIR light is narrowed, the local temperature can still exceed 100 °C even if the distance between the light source and the actuator is increased to 35 mm.
As a result, the jumping and rolling behaviors of the hydrogel can be actuated by the NIR light when the distance between the light source (equipped with a collimator) and the actuator is 35 mm, as shown in new Supplementary Figure 31. Therefore, the actuation distance could be increased from < 4 mm before narrowing the divergent NIR light to ~35 mm after narrowing the NIR beam. It is anticipated that the actuation distance could be further increased when utilizing a NIR light source with higher power.
ii) There are some differences between light-assisted technology and manual operation, for example: a) the light-assisted technology is an untethered propelling method suitable for actuating the hydrogel actuator in a small and enclosed space (Mobile Microrobotics, MIT Press, Cambridge, MA, 2017) which cannot be achieved by the manual operation; b) the position of the laser light source could be precisely controlled by mechanical method through the utilization of a XYZ translation stage which is difficult to achieve for the manual operation method.
We have thus added these new supplementary figures as Supplementary Figures 28, S30 and S31 and the corresponding descriptions are added on page 15, line 1 of the second paragraph: "In addition, we have studied the actuation of the current hydrogel actuator from a longer operation distance and the actuation of bigger sized ones, which may facilitate its further application in the practical field. To this end, it is desirable to keep the irradiation 10 intensity by narrowing the divergent NIR beam through the application of a collimator at the end of the NIR optical fiber (Supplementary Figure 28). The NIR laser beam without a collimator is divergent so that the spot area on the hydrogel actuator is dependent on the distance from the laser to the object and the temperature of the hydrogel actuator is only higher than 100 °C when the distance between the NIR light and the actuator is less than 4 mm (with a spot area of approximately 0.28 mm 2 at 2 mm separation) as shown in

'Also, energy density and power density should be characterized.'
Response: The light actuated shape deformation caused by the bubble formation due to the photothermal effect of IONPs produces energy (Eproduced) that generates jumping (Ejumping) or rolling (Erolling) motion by overcoming the stiction force to the ground (Estiction), the air drag (Eair drag), the internal energy dissipation (Edissipation), or the friction between the actuator and the ground (Efriction), as shown in the following (Soft Matter 2010, 6, 4342): In case of jumping:

12
The minimum produced energy could be estimated by Ejumping (in case of jumping) or Erolling (in the case of rolling) because the produced energy (Eproduced) is higher than that of jumping (Ejumping) or rolling (Erolling).
Among others, Ejumping can be analyzed based on the Newtonian equation as follows: where m is the mass of the hydrogel actuator which is 7.95 mg; vj is the take-off velocity. The energy density of the hydrogel actuator can thus be estimated by the following equation:

Energy density = Ejumping / mexpanding
where mexpanding is the mass of the actuation part of the actuator (~0.21 mg), which is calculated from the size of the local expanding part (mexpanding = ρ × Vexpanding, where ρ is the density of the hydrogel actuator; Vexpanding is the volume of the local expanding part). On the other hand, energy density and power density of the rolling behavior can be calculated based the following equations, respectively: 13 Erolling = 1/2 × mvr 2 Energy density = Erolling / mexpanding = 1/2 × mvr 2 / mexpanding Power density = Energy density / tr = 1/2 × mvr 2 / (mexpanding × tr) where tr and vr are the actuation time and velocity, respectively. mexpanding is the mass of the actuation part of the actuator (~0.21 mg), which is calculated from the size of the local expanding part (mexpanding = ρ × Vexpanding, where ρ is the density of the hydrogel actuator; Vexpanding is the volume of the local expanding part). Note that vr is dependent on the laser In case of jumping:

Eproduced = Erolling + Estiction + Eair drag + Efriction + Edissipation
The minimum produced energy could be estimated by Ejumping (in case of jumping) or Erolling (in the case of rolling) because the produced energy (Eproduced) is higher than that of jumping 14 (Ejumping) or rolling (Erolling). Among others, Ejumping can be analyzed based on the Newtonian equation as follows: where m is the mass of the hydrogel actuator which is 7.95 mg; vj is the take-off velocity. The energy density of the hydrogel actuator can thus be estimated by the following equation:

Energy density = Ejumping / mexpanding
where mexpanding is the mass of the actuation part of the actuator (~0.21 mg), which is calculated from the size of the local expanding part (mexpanding = ρ × Vexpanding, where ρ is the density of the hydrogel actuator; Vexpanding is the volume of the local expanding part). In addition, we have studied the jumping trajectory and destination by changing α. As shown in new Supplementary Figure 23(a, b), the jumping trajectories and destinations are tunable by adjusting α through varying the irradiation position. Note that the irradiation is fixed at 0.39 W.
Furthermore, the jumping trajectory and destination of the actuator could also be adjusted by changing the laser power (α is fixed at 70°), as can be seen from new Supplementary   Figure 23(c, d).
These experiments indicate that the jumping trajectories and destinations of the hydrogel actuator are controllable by adjusting the laser power and the light irradiation position.
Among others, the irradiation position and the laser power determine the take-off angle and the jumping height, respectively.
For rolling, the NIR light is controlled to irradiate the side part of the actuator with 0.14 W light irradiation and α = 60°, as illustrated in the following figure: Furthermore, we have studied the rolling destination by changing α and laser power. As shown in the following figure (a-b), the rolling destinations are controllable by adjusting α through varying the irradiation position (at fixed laser power, 0.14 W) or the laser power (at fixed α = 60°): Rolling destination of the hydrogel actuator when changing (a) α (laser power is fixed at 0.14 W) or (b) the laser power (α is fixed at 60°). 19 The above experiments indicate that the actuator's trajectories and destinations can be controlled for both jumping and rolling behaviors when the irradiation position and laser power are fixed. And different trajectories and destination of the actuator could be realized by tuning the irradiation position or the laser power.
We have thus added new Supplementary Figures 20-24  Then the laser is moved to position B to achieve the second-time actuation (jumping followed by rolling from position B to position C) in Supplementary Figure 25(b).
Despite this controllability, we agree with the reviewer's comments that it's challenging for this hydrogel actuator to achieve the continuous actuation, which may be solved by the following: Directly mount a laser source on the hydrogel actuator through proper design so that the laser could move with the hydrogel actuator together, which may lead to the continuous actuation of the hydrogel actuator. Note that the capability to mount a battery 24 containing detectaphone on the hydrogel actuator has been demonstrated in Supplementary   Figure 26(b), implying the possibility to directly mount a NIR laser fiber (if not a laser source) on the actuator. Work along this direction is currently under progress. And in this paper, we mainly report an in-air hydrogel actuator with fast response and superior moving performance, which steps away from the water environment required for the previously reported hydrogel actuators.
We have thus added this figure as Supplementary Figure  Response: In the current study, the Young's modulus of the hydrogel is measured by atomic force microscopy (AFM), as illustrated in the following figure: Typical force curve of the hydrogel actuator containing IONPs obtained by AFM.
The Young's modulus in this case is 108 MPa. However, we also believe that the reviewer's opinion is correct. We have synthesized the stripe-shaped hydrogel and measured its stress-strain curve by utilizing the Instron tensile tester, as shown in the following figure: 26 Tensile stress-strain curve of the hydrogel with 2 wt% Fe3O4 content.
The Young's modulus in this case is calculated to be about 1.34 MPa, which is different from that obtained based on AFM measurement. Comparing the above two results, we think the Young's modulus measured by the tensile tester is more accurate due to the following reason: The tensile tester measures the Young's modulus of the whole material, while AFM only measures the modulus of a small and localized surface. However, for the AFM measurement in the current case, the modulus of the hydrogel surface may be significantly higher than that of the bulk material because: During the AFM measurement, a NIR laser beam is applied which illuminates both the AFM tip and the IONP containing hydrogel. The NIR irradiation may cause the photothermal effect of IONP, leading to the dehydration of the pristine IONP containing hydrogel, the hardening of its surface and the increase of the Young's modulus.
Therefore, for accuracy, we adopt the tensile tester result instead of the AFM result. And the Young's modulus of the hydrogel with different IONP content measured by the tensile tester is shown in the figure below: 27 Young's modulus of the hydrogel with different IONP content.
We have thus added this figure as Supplementary Figures 11c and 15  ii) Furthermore, we have synthesized the hydrogel comprising the actuator (containing 2 wt% IONP) in the shape of a stripe and studied its mechanical property by utilizing the Instron tensile tester. As can be seen from Supplementary Figure 11, the fracture of the hydrogel containing 2 wt% IONP occurs when the strain is higher than 42 %. According to the numerical simulation (the same simulation shown in Figure 2o in the manuscript), the maximum strain of the hydrogel actuator during bubble formation is estimated to be around 20%, which is much lower than the fracture strain of the hydrogel. Therefore, the hydrogel does not fracture during the bubble formation process, which also accounts for the reusability of the hydrogel actuator (more than 100 times repetitive actuation, Supplementary   Figure 11).
We have thus added this figure as Supplementary Figures 11 and 12 and the corresponding descriptions are added on page 9, line 6 of the second paragraph: "According to the numerical simulation (Figure 2o), the maximum strain of the hydrogel actuator during bubble formation is estimated to be around 20%. This strain is lower than the fracture strain of the hydrogel comprising the actuator (i.e., > 42%, Supplementary Figure 11). This, together with the negligible changes in the surface morphology of the hydrogel comprising the actuator before and after NIR actuation (Supplementary Figure 12) and negligible water loss after the motion process (Figure 2m), indicates the hydrogel does not fracture during the bubble formation process, which also accounts for the re-usability of the hydrogel actuator (more than 100 times repetitive actuation, Supplementary Figure 13 We hope that our responses and the revised manuscript are satisfactory to you and the reviewers. We will be happy to provide any further information if needed. Authors deeply revised the work following the reviewers' comments and remarks. They added interesting new explanations and technical details in the text, thus improving the quality of the manuscript. Reviewer #2 (Remarks to the Author): The authors have addressed my major concerns. Publication in NC of the paper is recommended. The authors may consider the following possible minor improvements or clarifications, which are not mandatory: For the first question that I raised.
From Equation 2-3, I understand the relation between contact area and temperature. But what is the relation between contact pressure and bubble size? Could you directly give the analytic solution or simulated results of scaling relation between F-jumping and actuator size? Why in Figure S9, the distance from the contact point between the hydrogel actuator and substrate was set as the x-axis for contact pressure, as we only care the initial time point when distance between hydrogel and substrate = 0? I think the distance between the light source and the actuator as X-axis would make better sense, isn't it?
For the second question that I raised.
"in a small and enclosed space (Mobile Microrobotics, MIT Press, Cambridge, MA, 2017)" Could you give more practical & specific examples, that can manifest the advantages of your light assisted technology (distance ~35mm)?
For the third question that I raised.
I appreciate the authors analyzed the energy density from the "effect" side, that is, to estimate how much kinetic energy has been generated. It would be more informative if the authors can provide some energy analysis in terms of efficiency. For example, how much heat energy supplied by the light source, and the strain energy due to the bubble generation and hydrogel deformation. This would be helpful for the future optimization and provide design guidance for readers.