All-optical driven soft crawler with complex motion capabilities

ABSTRACT

The emergence of millimeter-scale soft actuators has significantly expanded the potential applications in areas such as search and rescue, drug delivery, and human assistance, due to their high flexibility. Despite these advancements, achieving precise control over the intricate movements of soft crawlers poses a significant challenge. In this study, we have developed an all-optical approach that enables manipulation of propulsive forces by simultaneously modifying the magnitude and direction of friction forces, thereby enabling complex motions of soft actuators. Importantly, the approach is not constrained by specific actuator shapes, and theoretically, any elongated photothermal actuator can be employed. The actuator was designed with an isosceles trapezoid shape, featuring a top width of 2 mm, a bottom width of 4 mm, and a length of 8 mm. Through our manipulation approach, we showcase a proof-of-concept for complex soft robotic motions, including crawling (achieving speeds of up to 2.25 body lengths per minute), turning, avoiding obstacles, handling and transferring objects approximately twice its own weight, and navigating narrow spaces along programmed paths. Our results showcase this all-optical manipulation approach as a promising, yet unexplored tool for the precision and wireless control for the development of advanced soft actuators.

GRAPHICAL ABSTRACT

KEYWORDS:

• Light driven soft actuator
• all-optical manipulation
• laser manipulation

1. Introduction

Soft actuators have attracted extensive interest and intensive research due to their superiority in simulating biological behaviors [1–10]. They are expected to perform tasks in complex environments and interact safely with humans due to their soft and deformable bodies [11,12]. Compared with rigid structures, soft actuators have greater degrees of freedom [13–15], are more flexible in action, and can perform specific tasks [16–21]. However, current soft robots, compared with their rigid counterparts, often exhibit less precision and poorer autonomous performance. These differences are attributable to a lack of precisely controlling of the deformation of soft actuators. Precise control of the direction and extent of the deformation of soft materials is more difficult than that of rigid materials.

It is important to develop suitable schemes that can precisely modify the deformation of soft materials and further control the movement of soft robots. The current control method of soft robots can be summarized as the conversion of external stimuli into forward propulsion such as light [22–24], electricity [25–27], magnetic fields [28–30], pH [31], air pressure [32,33] and humidity [34,35]. Especially for electro-thermally driven actuators, the effective electrothermal effect endows small flexible actuators with excellent mobility. At the same time, flexible actuators utilizing the electrothermal effect can achieve transparency, which significantly broadens the application fields of flexible actuators. In this process, researchers have explored various approaches to generate mismatch in friction forces and create forward propulsion. These include the addition of anisotropic structures [36–39] or using multilegged structures to independently stimulate specific responsive regions for crawling [40–44]. However, there are some urgent issues that need to be addressed for the current movement control of crawling actuators. Soft crawlers with anisotropic structures have demonstrated the ability to achieve efficient forward locomotion. But the flexibility of these methods is insufficient to meet the demands of multitasking in soft crawlers. Light driven actuators, compared with actuators driven by other stimuli, have advantage of high driving precision and easy to operate. The presented optical driven actuator leverages the precision of laser targeting to initiate movement, allowing complex motions such as crawling and steering by selectively illuminating specific parts of the actuator. Meanwhile, light driven actuators do not need large equipment to generate controlling field such as magnetic fields, vibration environments, or electric fields. This phenomenon enhancing the feasibility of deploying optical driven actuators in a wide range of settings, including those where space and portability are critical constraints. Compared to other light-driven micro-scale soft robots [18,45–52], the presented soft actuator has the ability to perform complex crawling and steering motions. Unlike traditional soft actuators that depend on material integration for motion, our design employs targeted laser light stimulation on specific parts of the actuator. This method allows for precise and modular control over its movements, facilitating complex maneuvers such as precise steering, without the need for intricate material patterning or integration. The actuator’s motion is achieved by directly stimulating selected regions with laser light, enabling real-time, on-demand adjustments. This approach offers a higher degree of movement control and precision, significantly enhancing its functionality for applications requiring intricate navigation and positioning.

Actuators driven by laser is regarded as a promising solution as the material can be selectively stimulated by the light spot [53,54], which means it can independently stimulate a specific part of a homogeneous material to achieve asymmetric deformation. Based on this phenomenon, it is possible to develop nondestructive light-driven schemes that can precisely control the deformation of actuator to perform multi specific tasks. In this study, we present a straightforward approach to control the movement of light-driven actuators by moving the laser spot. Both the direction and magnitude of the friction force between the actuator and the ground can be manipulated by scanning the laser spot in certain trails. In addition to the control method, flexible materials with good photothermal effect are also one of the necessary conditions to realize the efficient movement of actuators. Carbon nanotubes (CNTs)/polydimethylsiloxane (PDMS) composite films, which exhibit reversible actinic bending, were used in the study. Its unique optical-thermal property makes carbon-based bilayer actuators highly valuable in applications such as artificial muscles and biomimetic robotics, where precise and reversible actuation is necessary. Based on the above materials and inspired by the crawling mechanism of tea inchworm, a soft crawling actuator is presented (isosceles trapezoid, head width: 4 mm, tail width: 2 mm, length: 8 mm), as shown in Figure 1a. The movement of the proposed actuator is achieved through patterned cutting of CNTs/PDMS bilayers using a femtosecond laser, with subsequent driving provided by near-infrared (NIR) laser irradiation. During the actuation process, only the head and tail of the actuator make contact with the ground, while the body undergoes alternating bending and straightening motions, propelling the actuator forward. The actuator can transform from curved state to a near-flat state in a short period of time (approximately 2 seconds) under the stimuli of NIR laser. Once the NIR laser is removed, the actuator quickly returns to its initial curved state and resumes forward motion within a few seconds (around 3 seconds), demonstrating remarkable response performance. Benefitted from the soft body and the driving strategy, the actuator is able to traverse various terrains, including slopes, pipes, slits, and high humidity environments (shown in Figure 1c). Additionally, the actuator exhibits functionalities such as obstacle avoidance and item transportation, showcasing its excellent adaptability to diverse applications.

Figure 1. Inchworm-like soft crawling actuator based on CNTs/PDMS bilayer film and its fabrication method. a) an inchworm-like soft crawling actuator that can adapt to various terrains. b) the simplified preparation process of CNTs/PDMS bilayer actuator. c) schematic diagram of the passing through a slit, turning, and passing through a tunnel of the actuator.

2. Results and discussions

2.1. Fabrication and properties of CNTs/PDMS

The fabrication process of the CNTs/PDMS bilayers is conducted as follows. First, a suspension of CNTs is spin coated onto a cover glass to form a uniformly distributed CNTs film (with a thickness of ~9 µm). Subsequently, uncured PDMS is drop casted onto the CNTs film and cured at a temperature of 100°C, achieving a thickness of ~116 µm). This process culminates in the creation of the CNTs/PDMS bilayered film. Carbon nanotubes, which are cylindrical structures composed of carbon atoms, boast exceptional mechanical, electrical, and thermal properties. Their extraordinarily high thermal conductivity makes them ideal for creating thermal conduction networks. Since the coefficient of thermal expansion of PDMS (300 ~ 310 ppm/K) is higher than that of CNTs (1.5 ppm/K), thermal stress occurs at the interface between the two layers. Continuous irradiation causes the side with a higher coefficient of thermal expansion to bend toward the side with a lower coefficient of thermal expansion. After peeling off the bilayered film from the cover glass, the release of thermal stress causes the film to bend toward the PDMS side at room temperature (29℃), as depicted in Figure 1b and Figure S1. By utilizing a femtosecond processing system, soft bilayered actuators with arbitrary shapes can be created through direct cutting of the bilayered film, ensuring minimal thermal impact and enabling extremely precise cutting due to the ultra-short pulse duration of the femtosecond laser. The CNTs layer typically exhibits a higher absorption efficiency in the near-infrared region, with a wavelength of 808 nm serving as the driving source for the actuator to effectively induce the photothermal effect of CNTs, thus facilitating the actuator to generate significant deformation or deflection. Simultaneously, this laser provides enhanced manipulation flexibility due to its small spot size and high optical power density. By adjusting the laser power and focusing position, precise control over the location and intensity of laser irradiation can be achieved. When exposed to NIR laser irradiation in room temperature, the heat absorbed in irradiated area increases (The highest local radiation temperature reaches around 110℃). As a result of the thermal stress-induced strain arising from the difference in the coefficient of thermal expansion between the PDMS and CNTs layers, the actuator bends toward the side of the CNTs layer. Upon turning off the NIR laser, the actuator gradually cools down to room temperature, causing it to return to its initial state. In addition to the material’s intrinsic properties, the degree of the actuator’s deformation is intimately linked to temperature. The actuator contracts more when starting from a lower environmental temperature due to the effect of temperature differences. This leads to a longer duration of maximum deformation under illumination and a quicker return to its original state once the irradiation ceases. Throughout this process, the deformations experienced by the actuator are reversible and repeatable.

In order to investigate the properties of the CNTs/PDMS bilayered film, a rectangular-shaped actuator (13 mm × 2 mm) was fabricated. One end of the actuator was fixed by tweezers. The initial state of the actuator, observed at room temperature without NIR laser irradiation, exhibited a slight bending toward the PDMS film side, as depicted in Figure 2a and (i) in Figure 2b. Visually, starting from the forceps-clamped end, the total free end of the actuator is divided into three equal-length segments. The starting point of laser irradiation for each segment, with deformation occurring only in the irradiated part. This design allows the actuator to mimic the movement of a finger, with two digital joints, as illustrated in Figure 2(a,b), and Movie S1.

Figure 2. Properties of CNTs/PDMS with NIR irradiation. a) b) the irradiated area of the rectangular actuator made of CNTs/PDMS will locally bend to the CNTs side to imitate the bending of human finger joints when different parts of it are exposed to NIR. Curves of the bending angle c) and temperature d) of the rectangular actuator with NIR irradiation (100 mW, 150 mW, and 200 mW) 5 s and after NIR irradiation was turned off. e) the maximum bending angle and displacement on the x-axis and y-axis of the rectangular actuator with NIR irradiation (30 mW, 50 mW, 70 mW, 100 mW, 120 mW, 150 mW, 170 mW, 200 mW, 220 mW, 250 mW, and 270 mW). f) fatigue tests of rectangular actuator under NIR irradiation (120 mW).

To quantitatively investigate the thermal response of the soft actuator, the bending angle β and temperature were measured when irradiated by NIR laser with power ranging from 100 mW to 300 mW. The β is defined as the angle between the line connecting the midpoint of the actuator’s fixed end to the midpoint of its free end and the horizontal line. The laser source remained fixed at a distance of 10 cm from the actuator, targeting the area close to the fixed end. A thermal infrared imager was utilized to record the bending angle and temperature. The experimental results demonstrated a positive correlation between the bending angle and the power of NIR laser. As the power of NIR laser increased, the bending angle of the actuator gradually increased. The rate of increase was the highest during the first 2 s of irradiation and then gradually slowed down, stabilizing in about 5 seconds. The bending angles at 0 s, 2 s, and 5 s were measured as 38.5 degrees, 55.8 degrees, and 84.5 degrees, respectively (as shown in Figure 2c). When the laser was turned off, it was observed that the bending angle of the actuator rapidly decreased within 1 second and gradually returned to its initial state within the subsequent 0.5 seconds. The deformation characteristics of the CNTs/PDMS bilayered film are significantly influenced by its thickness. Drawing on prior experimental research [55] and specific experimental applications, the optimal thickness is approximately 125 µm. An overly thick PDMS may increase the bilayer’s overall stiffness, thereby limiting its flexibility and bending capabilities. On the other hand, a PDMS layer that is too thin may reduce the bilayer’s strength, increasing the risk of rupture.

Moreover, the temperature change of the irradiated area exhibited a positive correlation with the power of the NIR laser. During the experiment, the actuator was first irradiated for 5 seconds and then allowed to recover to its initial state by turning off the laser. Within the first 3 seconds of irradiation, the temperature of the irradiated area increased sharply, with a growth rate exceeding 90%. Subsequently, the temperature slowly increased and stabilized over the following 2 s. The recorded temperatures at 1 s, 2 s, and 5 s were 70.8°C, 89.8°C, and 105.6°C, respectively. When turning off the NIR laser, the temperature of the irradiated area dropped rapidly, with a decay rate of over 90% within 4 s. After 5 s of laser deactivation, the temperature of the irradiated area almost returned to its initial level (as depicted in Figure 2d). Furthermore, when a NIR laser with a power of 250 mW was employed in the experiment, the average temperature of the irradiated area was measured as 108.4°C, indicating that the local temperature did not significantly increase with higher NIR power.

Additionally, to further evaluate the influence of laser power on the performance of the actuator, the horizontal displacement and vertical displacement of the actuator were measured, as illustrated in Figure 2e (The point ‘o’ is the clamping point of the gripper, point ‘b’ is the midpoint of the actuator after bending, the distance from point ‘o’ to ‘c’ is the length of the actuator, point ‘a’ to ‘c’ is the horizontal displacement of the actuator, point ‘a’ to ‘b’ is the vertical displacement of the actuator). The laser intensities were set at 30 mW, 50 mW, 70 mW, 100 mW, 120 mW, 150 mW, 170 mW, 200 mW, 220 mW, 250 mW, and 270 mW. While the bending angle, horizontal displacement, and vertical displacement of the actuator were measured separately when the actuator reached a stabilized state. The experimental results are presented as follows (Figure 2e and Table 1):

Table 1. The influence of gradient laser power on the performance of the actuator.

Laser power (mW)	30	50	70	100	120	150	170	200	220	250	270
Bending angels (degrees)	2	4	15	61	70	79	84	92	93	95	100
Vertical displacements (mm)	0.13	0.49	1.52	10.33	12.11	13.93	14.59	15.9	16.06	16.36	16.88
Horizontal displacements (mm)	0.27	0.53	2.52	4.61	3.97	3.26	2.53	1.47	1.37	0.92	−0.01

It is observed that the bending angle, vertical displacement, and horizontal displacement of the actuator exhibited a rapid increase when the NIR power rising from 50 mW to 100 mW. The increasing rate of bending angle and vertical displacement gradually slowed down when the laser power rising to the range between 100 mW and 200 mW, while the horizontal displacement showed a decreasing trend. This behavior can be explained as follows. The actuator firstly reached to a state parallel to the horizontal plane at approximately 100 mW, which means further increases in laser power will cause an overbend to result in downward bending. Beyond 200 mW, the maximum bending angle and vertical displacement did not show significant increases, which is similar to the temperature response of the actuator at high NIR power. Thus, the horizontal displacement continued to decrease and eventually approached its initial value.

Fatigue resistance is an important parameter for assessing the application potential of soft actuators. When the area near the fixed end of the actuator was repeatedly irradiated with 120 mW NIR laser for 100 cycles, the maximum bending angle consistently remained above 40°, indicating good durability and suitability for practical applications (Figure 2f). It is evident that when the NIR power exceeded 200 mW, neither the local temperature nor the bending angle of the actuator showed significant increases, suggesting that 200 mW may be the optimal choice. Moreover, the study of horizontal displacement and vertical displacement demonstrated that the actuator is capable of a wide range of response deformations relative to its own length, highlighting its versatility.

2.2. The control strategy of a soft crawling actuator made of CNTs/PDMS

In order to explore the potential applications of the soft actuator, a crawling actuator based on CNTs/PDMS was fabricated. To achieve different contact areas between the actuator’s head and tail with the ground, the actuator was designed with an isosceles trapezoid shape. It features a top of 2 mm, a bottom of 4 mm, and a length of 8 mm. The trapezoid bottom was chosen as the head of the actuator, while the trapezoid top served as the tail.

In its initial state, the soft actuator bent toward the PDMS side in room temperature. When an NIR laser (200 mW) was applied to scan from the head to the tail of the actuator, the head was first pressed down and gradually extended forward, followed by the gradual extension of the tail toward the backward direction as the NIR laser moved toward the tail. When the NIR laser was turned off, both the head and tail of the actuator retracted inward. The larger contact area between the actuator head and the ground results in a greater frictional force at the head compared to the tail. Therefore, the tail retracted more than the head, propelling the soft actuator forward in the direction of the head (Figure 3a,b). The gray dashed lines in Figure 3b indicate the crawler’s appearance in the previous stage. The crawling process of the crawler was analyzed, focusing on the changes in the contact angle between the crawler and the ground during NIR laser irradiation (Figure 3c). In the initial stage (i), when the head was irradiated by laser, it underwent deformation, causing a backward shift in the crawler’s center of gravity and resulting in morphological changes. The contact angle between the head and the ground decreased, and the tension angle in the head shifted toward the upper right, leading to a decrease in downward pressure and a reduction in frictional force. Concurrently, the contact angle between the tail and the ground increased, and the tension angle shifted toward the lower left, resulting in increased downward pressure and enhanced frictional force in the tail. As the light spot moved toward the body in stage (ii), the entire crawler started to extend. Due to decreased frictional force in the head, it experienced displacement and moved forward. When the light spot reached the tail in stage (iii), the crawler reached its maximum extension. At this point, the NIR laser was turned off to reach stage (iv). Although the contact angle between the head and the ground remained acute, the crawler underwent contraction, leading to downward pressure and upward tension in the head. Consequently, the frictional force in the head surpassed that in the tail, causing the head to remain relatively stable while the tail moved toward the head, entering stages (v) and (vi).

Figure 3. Physical diagram a), schematic diagram b), and force analysis c) of the crawling actuator at different stages in a drive cycle with NIR irradiation.

A detailed analysis of a single crawl cycle was conducted to further understand the crawling mechanism of the actuator. In the initial state, the projected length of the actuator on the ground was approximately 5.5 mm. When the NIR laser irradiated the head of the actuator, the head began to press down and slightly extend toward the front. As the NIR laser moved to the middle part of the actuator, it quickly extended toward both the head and tail, while maintaining downward pressure on the entire body. At this stage, the extension distance of the tail was noticeably greater than that of the head, due to the relatively lower resistance for rearward extension.

As the NIR laser gradually moved toward the tail of the actuator, the downward pressure on the tail increased, while the head was released and slowly began to recover on a local scale. The resistance from the ground against the tail increased, preventing further extension of the tail and pushing the head forward substantially. After approximately 1.7 seconds, the NIR laser was removed from the actuator. At this point, the head had extended forward by nearly 2 mm, which was approximately 30% of the projected length in the initial state. Subsequently, the head and tail of the crawling actuator gradually started to shrink inward. Due to the greater friction between the head and the ground compared to the tail, the entire body moved forward. After approximately 3.2 seconds, the actuator returned to its initial state. At this time, the head had moved forward by 1.7 mm, which accounted for approximately 20% of its own length (Figure 4a). To test working life of the crawling actuator, we conducted 50 cycles of light irradiation experiments (Movie S2). Furthermore, we tested the laser irradiation stimulation for 400 cycles, and the soft actuator did not show any damage, which demonstrates the stability and long lifespan of the actuator.

Figure 4. Motion characteristics of the soft crawling actuator when NIR drives it with different irradiation strategies. a) changes in the soft crawling actuator body length and the traveling distance of the head and tail when irradiating from its head to tail with NIR (200 mW). Trajectory of the soft crawling actuator and its displacement on the x- and y-axes. When NIR drives it to go straight b), turn left c), and turn right d) with different irradiation strategies. The insets in (b-d) are physical diagrams and thermal distribution diagrams of the crawler during the crawling process. (the scale bars are 5 mm).

In addition, as a comparison, the NIR laser was also irradiated from the tail to the head of the actuator. It was observed that the actuator still exhibited good crawling performance in this configuration, consistently moving forward with a displacement greater than 1 mm in one driving cycle (Figure S2).

To achieve different crawling trajectories, various illumination modes were designed considering the temperature and stress changes near the irradiated area. To make the actuator crawl forward in a straight line, the NIR laser was designed to follow the symmetry axis from the tail to the head of the actuator (Figure S3a). Although the actuator typically exhibited an asymmetrical bending trend, the scanning trajectory was adjusted to compensate for this asymmetry. With this configuration, the actuator crawled approximately 14 mm along the Y-axis (approximately 175% of its own length) and maintained a lateral displacement on the X-axis within 2 mm after driving for 120 seconds (Figure 4b, Figure S4 and Movie S3).

When the NIR laser was directed from the tail to the head along the right side of the actuator, the actuator steered to the left during crawling. The right side of the actuator experienced more stretching than the left side, leading to asymmetric stretching. During the subsequent contraction process, the differential driving force between the left and right sides caused the head to deflect to the left by an angle θ, resulting in a left turn of the actuator (Figure S3b). After driving for 120 seconds, the actuator moved forward approximately 14 mm along the Y-axis (approximately 175% of its own length) and shifted approximately 7 mm to the left on the X-axis, with a left deflection angle of approximately 26.6 degrees (Figure 4c, Figure S5 and Movie S4).

Similarly, the actuator can be steered to the right using a similar principle (Figure S3c). In this configuration, the actuator moved forward approximately 14 mm along the Y-axis (approximately 175% of its own length) after driving for 120 seconds, and shifted approximately 6.5 mm to the right on the X-axis, with a right deflection angle of approximately 23.9 degrees (Figure 4d, Figure S6 and Movie S5).

The soft crawling actuator demonstrates its versatility by being able to perform various tasks in different applications using different driving strategies. By scanning the NIR laser along the symmetry axis from the tail toward the head of the actuator, it can crawl a distance twice its projected length on a 10-degree inclined plane in 67 seconds. Compared to crawling on flat ground, the climbing actuator needs to overcome the effects of gravity on the inclined surface. The single-cycle illumination involves four steps: When the laser illuminates the front half of the actuator, the actuator’s head begins to extend forward; At the middle, the extent of head extension reaches its maximum; Subsequently, when the laser irradiates the rear half, the tail begins to extend; Finally, when the laser scans the tail, the entire actuator begins to contract. Under the influence of asymmetric friction at the head and tail, the actuator moves forward a certain distance (As shown in Figure S7). Crawling a distance of 5 times its own length in a horizontally placed pipe with a diameter of 7 mm. The pipe, made of transparent acrylic material, boasts a high light transmittance of up to 93%. However, its non-planar characteristics lead to a reduction in the laser beam’s energy density. To counteract this loss, the power of the near-infrared light was adjusted to 220 mW. Within 184 seconds, a ‘C’-shaped motion trajectory was designed to simulate the crawling and obstacle avoidance of the actuator. When the actuator is about to climb to a corner where a change in direction is needed, the direction of the laser irradiation is altered to achieve redirection. (Figure 5a-c and Movies S6-S8).

Figure 5. Application of a soft crawling actuator in different working conditions. a) under the irradiation of 200 mW NIR, the soft crawling actuator can crawl on a slope of 10 degrees. b) passing through a tunnel with a diameter of 7 mm. c) crawling along the C-shaped trajectory to avoid obstacles. d) crawling in a high humidity environment with continuous humidification. e) climbing over the ladder. f) passing through a slit with a height of only 1 mm in 146 s. g) transporting the object forward 5 mm in 26 s. (the scale bars are 5 mm).

The crawling speed of the actuator depends on the duration of a single laser scanning cycle and the time interval between each scanning cycle. When the scanning cycle and time interval were set at 8 s and 5 s, the actuator crawled 4 mm in 66 seconds, resulting in a speed of 0.06 mm/s. When the scanning cycle and time interval were set at 3 s and 2 s, the actuator crawled 9 mm in 30 seconds, achieving a speed of 0.3 mm/s (2.25 body length per minute), as shown in videos (Movie S9 and Movie S10).

The actuator also exhibits mobility in high-humidity environments, being able to crawl forward one body-length distance in a high-humidity environment within 21 seconds (Figure 5d and Movie S11). In unexpected, such as when dropped from a height and landing bottom up with its back side in contact with the ground, the actuator can still move in a ‘snake-like’ mode under NIR laser stimulation (Figure 5e, Figure S8 and Movie S12). Furthermore, the actuator can smoothly pass through narrow slits with a height that is only 1/3 of its own height. By irradiating the head, the local temperature increases, promoting the head to smoothly enter the slit and completely pass through it within approximately 148 seconds (Figure 5f and Movie S13). The actuator is also capable of transporting objects using this driving strategy. In an experiment employing a thin, folded aluminum sheet as the object, the masses of the actuator and aluminum foil are weighed 4.8 and 9.3 mg, respectively. The actuator crawled forward 5 mm during scanning the NIR laser for 26 seconds. This demonstrates that the actuator can generate a force approximately double its weight. (Figure 5g and Movie S14). These results demonstrate the adaptability and potential applications of the soft crawling actuator in various scenarios and tasks.

As a proof of concept, the response of the actuator in different temperature environments was investigated to understand its behavior in extreme conditions. Tests were conducted in environments with temperatures ranging from −8°C to 80°C. The results showed that the time required for the actuator to fully extend under NIR irradiation decreased as the ambient temperature increased. For example, at −8°C, the actuator took approximately 5 seconds to fully extend, while at 80°C, it only took approximately 2.2 seconds. This can be attributed to the actuator being influenced by the ambient heat. As the ambient temperature increases, the actuator experiences some degree of extension in its initial state. Consequently, the time required for full extension gradually decreases during NIR irradiation. On the other hand, when the ambient temperature increased from −8°C to 10°C, the shrinkage time of the actuator gradually increased from approximately 8.3 seconds to approximately 10 seconds. This could be due to faster temperature loss at significantly low temperatures, leading to quicker cooling of the actuator and faster shrinkage rates. However, as the ambient temperature increased from 10°C to 80°C, the shrinkage time exhibited a decreasing trend, reducing from approximately 10 seconds to approximately 5.4 seconds. This behavior is also influenced by the ambient heat. With higher ambient temperatures, the actuator already undergoes a certain degree of extension in its initial state. Consequently, the time required to restore the actuator to its initial state gradually decreases after the NIR laser is removed, as shown in Figure S9.

3. Conclusion

In summary, the all-optical driving approach enables a worm-like soft crawling actuator with several key features and capabilities. By controlling the irradiation area, power, and angle of the friction force, the actuator was able to effectively crawl forward, turn around, and crawl backward. Compared to actuators that rely solely on the mismatch of friction forces between the head and tail, this actuator utilized both the power and direction of the friction force, leading to more effective crawling directions. Additionally, by leveraging the local heating principle of the NIR-driven actuator, the actuator could perform turns with only the irradiation of the NIR laser. This capability allowed the actuator to exhibit steering, obstacle avoidance, and object transportation. Furthermore, the actuator demonstrated good adaptability to various environmental conditions, such as slopes, narrow passages, high humidity, low temperature, and high temperature. This adaptability opens up possibilities for the actuator’s application in different scenarios.

4. Experimental section/methods

4.1. Materials

Multiwalled carbon nanotubes (CNTs) were commercially obtained from Nanjing XFNANO Materials Tech; a PDMS mixture (Sylgard 184 silicone elastomer, from Dow Corning) was commercially purchased.

4.2. Femtosecond Laser Direct Writing Optical System

The laser beam (104 fs, 1 kHz, 1030 nm) from a regenerative amplified Ti: sapphire femtosecond laser system (Legend Elite-1K-HE, Coherent, U.S.A.) was employed for ablation. During the fabrication process, the laser beam was guided onto the sample via a galvanometric scanning system (SCANLAB, Germany), which made the laser beam focus and scan along the x/y coordinate direction.

4.3. NIR Laser

An NIR laser (808 nm, Anford, Shenzhen, China) was used to provide external stimuli for the soft actuator to bend and move.

4.4. Measurement of temperature, bending angle, and displacement

We used a thermal infrared imager (Fotric 225–1) to capture infrared thermal images and record the temperature variation during light-driven actuation. The mobile camera (PAR-AL00, HUAWEI, Shenzhen, China) and the mobile camera (EVA-AL10, HUAWEI, Shenzhen, China) were fixed in front and top of the actuator to record its bending angle and displacement.

4.5. Characterization

SEM images of CNTs/PDMS were obtained by a Carl Zeiss Gemini 300 field emission scanning electron microscope.

Acknowledgments

We acknowledge the Joint Laboratory of Smart Material Devices and Equipment at HFUT for the fabrication and measuring of samples. This work was partly carried out at the USTC Center for Micro and Nanoscale Research and Fabrication.

Supplemental Material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/19475411.2024.2353312.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by the National Natural Science Foundation of China [62105090, 22275048, 22411530048]; the Fundamental Research Funds for the Central Universities [JZ2023YQTD0074]; the National Key R&D Program of China [2021YFF0502700]; Anhui Provincial Natural Science Foundation [2008085J22]; the USTC Research Funds of the Double First-Class Initiative [YD2340002009].