Inspired by microorganisms in nature, artificial micromotors that propel themselves by converting diverse energies have been designed to perform complex tasks, such as cargo transport [1–3], advanced nanomanufacturing [4, 5], and environment remediation in microscopic environments [6–11]. Their motile behavior also further endows them great advantages as next-generation nanorobotics used for biomedicine [12, 13].

Mg-based micromotors, as a typical class of self-consuming artificial micromotors, not only could be powerfully driven by the Mg-H₂O reaction in physiological condition but also produce a byproduct of Mg²⁺ ions, an additional nutritional ingredient that benefits the human bodies [14, 15]. Additionally, the vast quantities of therapeutic hydrogen generated in the propulsion can further be used to eliminate excessive hydroxyl radicals which are harmful to living bodies [16]. They displayed superior compatibility for biological systems and have already been applied for drug delivery in the gastrointestinal tract of mice with enhanced therapeutic efficacies [17–22]. However, they generally suffer from poor motion controllability and relatively short motion lifetime stemming from the single driving reaction and self-consuming feature and cannot stop their movement unless the solid “fuel” Mg cores are almost consumed [23, 24]. To prolong the lifetime and improve the motion controllability of Mg-based micromotors, several strategies have so far been developed including tuning the opening size, controlling the pitting anions, and being functionalized with photocatalysts [21, 25], soft hydrogels [26], or biological materials [19, 25, 27, 28]. Yet they still face challenges in stimuli-responding propulsion and desirable “on/off” characteristics, and further efforts are necessary toward intelligence and controllability, which may lay the foundation for Mg-based micromotors to fulfill tasks in complex environments.

Here, we present a Mg-based micromotor with motion responsive to both H₂O₂ concentration and temperature in the surroundings. It consists of a Mg microsphere asymmetrically coated with Pt and temperature-sensitive poly(N-isopropylacrylamide) (PNIPAM) hydrogel layers in sequence. It can demonstrate the propulsion mechanism transformation between Mg-H₂O reaction and Pt-catalyzed H₂O₂ decomposition and unique concentration- or temperature-induced self-limitation behaviors like hovering. The temperature-induced volume phase transformation (swelling/shrinking) of PNIPAM and the H₂O₂ concentration both affect the permeability of H₂O₂ into the hydrogel layer and thus make the generation of O₂ bubbles in either single or bilateral sides, resulting in self-limitation motion. This work presents a stimuli-induced “hovering” strategy for the manipulation of Mg-based micromotors as well as a method for prolonging their lifetime by taking use of fuels from the surroundings. Given the possible changes of H₂O₂ concentration and temperature in human bodies, especially in diseased sites, the herein reported dual stimuli-responsiveness of Mg-based micromotors with extended motion lifetime may to some extent promise potential advanced biomedical applications.

Given the motion environment diversity of micromotors when used, we focus on the changes of chemical reaction propulsion mechanism and dual stimuli-responsive motion behaviors of Mg-based micromotors in the liquid environment system. The aim is to achieve their long-term lifetime and motion manipulation. In view of the popularity of temperature difference and low concentration H₂O₂ in either biological or environmental system, we have integrated Mg microspheres with catalytically active Pt to H₂O₂ and temperature-responsive poly(N-isopropylacrylamide) (PNIPAM) hydrogels to construct Mg-based micromotors. They were fabricated by asymmetrically coating Mg microspheres with Pt and PNIPAM hydrogel layers in sequence via magnetron sputtering and UV polymerization techniques according to our previous literature [29]. The as-fabricated Mg-based micromotors are about 40  μm in size and exhibit “opening” of about 14  μm of the surface (Figures S1 and S2). As shown in Figure 1, the Mg-based micromotor when operated in an aqueous solution (e.g., containing NaHCO₃ and poly(vinylpyrrolidone) (PVP)) is driven by asymmetrical H₂ bubble release generated by Mg-H₂O reaction due to the effective shielding of Pt layer from water in a gradually increasing confined space. When encountering catalyzed substrates such as H₂O₂, the propulsion mechanism of the Mg-based micromotor will transform spontaneously from the self-consuming Mg-H₂O reaction into Pt-catalyzed H₂O₂ decomposition. As H₂O₂ (\text{H}{{\text{O}}_2}^{-}+{\text{H}}_2\text{O}+2{\text{e}}^{-}⟶3\text{O}{\text{H}}^{-}, E^0=0.878\text{V} vs. SHE) has a stronger oxidization ability than H₂O (2{\text{H}}_2\text{O}+2{\text{e}}^{-}⟶{\text{H}}_2+2\text{O}{\text{H}}^{-}, E^0=-0.83\text{V} vs. SHE) [30], the reaction of Mg and H₂O₂ is much easier to occur than the Mg-H₂O reaction. As a result, O₂ bubbles from Pt-catalyzed H₂O₂ decomposition instead of H₂ bubbles propel the micromotor and the motion lifetime could be prolonged due to the transformation of driving reactions from the self-consuming mode to the catalytic mode. When the temperature is below the lower critical solution temperature (LCST, 32°C), the PNIPAM hydrogel is in the hydrophilic swelling state and has a large diffusion constant to solution [31]. Consequently, H₂O₂ aqueous solution can easily penetrate the gel layer to react with the middle Pt layer resulting in a temperature-induced self-limitation motion like hovering due to bubble O₂ recoils from bilateral sides. Here, the self-limitation motion behavior means that a micromotor moves in 20 s only within a circle with the radius less than its twice size. The area of the motion region is characterized as S_{20}=\pi \left\{{\left(x_{\mathrm{Max}}-x_0\right)}^2+{\left(y_{\mathrm{Max}}-y_0\right)}^2\right\}; here (x_{\mathrm{Max}}, y_{\mathrm{Max}}) is the farthest track point and (x_0, y_0) is the initial track point in 20 s. A similar transformation of motion behavior can also be triggered by increasing H₂O₂ concentration denoted as a concentration-induced self-limitation. A high concentration of H₂O₂ increases the concentration gradient between the two sides of the gel layer, inducing more H₂O₂ through the PNIPAM hydrogel layer for bilateral O₂ bubble generation. In a word, the as-constructed Mg-based micromotor can implement different motion behaviors stemming from the propulsion mechanism transformation when encountering Pt-catalyzed substrates such as H₂O₂ and respond to both the concentration of catalyzed substrates and temperature in aqueous environment with extended motion lifetime. This suggests that the as-prepared Mg-based micromotor may be taken as an example to address its intrinsic drawbacks of uncontrollable propulsion and limited motion lifetime arising from its self-consuming features. Although the H₂O₂ concentration used here is not in the nontoxic physiological range of less about 50  μM, the “fuel” concentration for the effective motor driving can be reduced via enhancing catalytic reaction activities, regulating motor structures, and so on [32, 33]. Moreover, the LCST of the hydrogel can be adjusted to fit the body temperature range by copolymerization [34, 35]. Thus, this will to some extent promote the development of Mg-based micromotors toward biomedical application.

Figure 2 and Video S1 show the motion behaviors of the Mg-based micromotor in the aqueous solution without or with 0.1 wt% H₂O₂ at 38°C or 22°C, respectively. All systems contain 0.5 M NaHCO₃ and 5 wt% PVP to peel off the Mg(OH)₂ passivation layer in Mg-H₂O reaction and to facilitate the suspension of the micromotor [36–38]. As shown in Figure 2(a), the Mg-based micromotor displays an obviously increased speed from 14.3  μm/s to 28.7  μm/s when moving from the aqueous solution without H₂O₂ (Figure 2(a), i) to that with 0.1 wt% H₂O₂ (Figure 2(a), ii) at 38°C. The speed increase is caused by evidently accelerating bubble formation and detachment, suggesting the possible driving mechanism transformation from Mg-H₂O reaction to Pt-catalyzed H₂O₂ decomposition. In the presence of H₂O₂, metal Mg prefers to react with H₂O₂ without H₂ production rather than with H₂O because of the higher reduction potential of H₂O₂ (0.878 V vs. SHE) than H₂O (-0.83 V vs. SHE) [30]. Simultaneously, Pt-catalyzed H₂O₂ decomposition has a much higher rate constant (about 3.60\times {10}^{-2}{\min }^{-1}) than Mg-H₂O reaction (about 1.05\times {10}^{-2}{\min }^{-1}) at an approximate pH of 7 [39, 40]. Additionally, the lower critical nucleation concentration of O₂ (68 mM) is much lower than that of H₂ (about 250 mM) bubbles [41, 42]. Both can induce vast quantities of O₂ bubble generation to powerfully drive the motor.

As the PNIPAM hydrogel is a temperature-responsive polymer with LCST, it usually exhibits a sharp volume phase transition at LCST, which is slowed by further increasing/decreasing temperature [43]. On the basis of the swelling-shrinking kinetics of the PNIPAM hydrogel, the collective diffusion coefficient for swelling (D_0=2.0\times {10}^{-7}\text{c}{\text{m}}^2{\text{s}}^{-1}) is about twice than that for shrinking (D_0=1.1\times {10}^{-7}\text{c}{\text{m}}^2{\text{s}}^{-1}) [44, 45]. Although at the temperature below the LCST (32°C) the PNIPAM has a volume swelling with a large cooperative (collective) diffusion of the gel network [31, 44], the micromotor can only be propelled by asymmetrical bubble recoils during the Mg-H₂O reaction because of the effective blocking from water by the middle Pt layer (Figure 2(a), i and iii). This will change when H₂O₂ appears in the system. As shown in Figure 2(a), iv, the micromotor shows an enhanced rotation with a decreased speed of about 15.8  μm/s and a decreased S_{20}, which looks like hovering because of bilateral bubble generation and release. This can be explained by the assumption that H₂O₂ aqueous solution may diffuse across the swelling hydrogel layer, reacting with the middle Pt layer to induce a self-limitation motion behavior of the Mg-based micromotor.

In order to explicate the temperature-induced self-limitation behavior, the corresponding angular velocity curves of Figure 2(a) are given in Figure 2(b). The single-side bubble-propelled micromotors (i, ii, and iii) show similar angular velocity curves with placid angular changes, while the bilateral-side bubble-propelled micromotor shows a relatively large angular change (iv). This can well explain the originality of the observed self-limitation state of the micromotor shown in Figure 2(a), iv. Under low Reynolds number conditions, the inertial effect can be neglected. The motion of the as-prepared Mg-based micromotor can be described by a set of differential equations for translation and rotation [46–48]:

where m represents the mass of the motor, J is the inertia tensor, and v and \omega denote the translation velocity and angular velocity. The terms that distinguish these Langevin equations from standard Langevin equations for inactive Brownian particles are the driving force and torque contributions arising from bubble detachment. Thereinto, F_{\text{Brownian}} and T_{\text{Brownian}} represent the diffusion arising from the Brownian force and torque with zero mean. The purely Brownian force for a passive particle can be described by a theoretical translational diffusion coefficient (D_T=k_{B}T/6\pi \eta R) [49] and a rotational diffusion coefficient ({\tau }_R^{-1}=k_{B}T/8\pi \eta R^3) [50]. In this system, as a spherical particle with a radius about 40  μm, the characteristic time scale for rotational diffusion {\tau }_R and translational diffusion coefficient D_T is calculated to be 728.5 min and 1.22\times {10}^{-10}\text{c}{\text{m}}^2/\text{s}, respectively. This means the effect of Brownian force on the motion of the as-prepared micromotor is limited. F_{\text{drive}} and T_{\text{drive}} are the reaction force and torque caused by the detachment of the bubbles, which will be balanced by the Stokes viscous drag force generated in the opposite direction during motion. Because of the spherical structure of the as-prepared micromotor, F_{\text{Stokes}}=6\pi \eta Rv (for translation) and T_{\text{Stokes}}=8\pi \eta R^3\omega, where R is the particle (hydrodynamic) radius and \eta is the fluid viscosity. For the motion of the as-prepared Mg-based micromotor, the contributions of F_{\text{Stokes}} and T_{\text{Stokes}} are passive depending on v and \omega and the rotation of bubble recoil-propelled micromotors mainly arises from the bubble detachment deviating from their symmetry [26, 51–53]. Therefore, a simplified force analysis of a micromotor based on the driving force is used to understand the motion states. F_1 and F_2 represent the instant propulsion forces provided by the bubble detachment from the opening and the coated side, respectively. For the cases shown in Figure 2(a), i, ii, and iii, F_2=0 and the micromotor is only driven by single-side bubble recoil on the opening. Their rotational motions are all caused by the rotational component of F_1 contributing an instant torque expressed as T_1=F_{1\perp }{L}_1=F_1{L}_1\sin \alpha, where F_{1\perp } represents the vertical component of F_1 for motor rotation, L_1 is the effective lever arm, and \alpha is the angle between the directions of the centroid connection line and F_{1\perp }. The differences in their motion states including translational and rotational velocities as well as rotational directions mainly depend on \alpha and the bubble detachment positions. For the micromotor moving in 0.1 wt% H₂O₂ aqueous solution at 22°C, neither F_1 nor F_2 is 0 due to bilateral bubble generation and ejection. As a result, the instantaneous torque that contributed to the micromotor’s rotation is the vector sum of the two torques: T_1\left(T_1=F_1{L}_1\sin \alpha \right) generated by bubble escape from the open end and T_2\left(T_2=F_2{L}_2\sin \beta \right) generated by bubble release from the coating side. When \alpha \beta >0, the micromotor exhibits an enhanced rotation corresponding to a large angular velocity (Figure 2(b), iv), while it shows a weakened rotation as (Figure S3, Video S2). The self-limitation behavior of the Mg-based micromotor in the H₂O₂ aqueous solution below the LCST of PNIPAM is attributed to the enhanced rotation caused by the bilateral-side bubble detachments.

The dissolved H₂ and O₂ in the aqueous solution before and after adding H₂O₂ are determined to verify the driving mechanism transformation from the self-consuming Mg-H₂O reaction to the Pt-catalyzed H₂O₂ decomposition. Figure 3 shows that when the motors move in the aqueous solution without H₂O₂, the dissolved H₂ in the solution increases gradually. This indicates that the driving reaction at this moment is a H₂-produced reaction as \text{Mg}+{\text{H}}_2\text{O}+2\text{HC}{\text{O}}_3^{-}\rightleftharpoons {\text{MgCO}}_3+{\text{CO}}_3^{2-}+2{\text{H}}_2. When H₂O₂ is introduced into the solution at 301 s, the dissolved H₂ drops sharply, while the dissolved O₂ rises from 2.7 mg/L to 19.9 mg/L. This implies that the driving reaction of the Mg-based micromotors is transformed to Pt-catalyzed H₂O₂ decomposition (producing O₂) instead of the Mg-H₂O reaction (producing H₂). The initial O₂ concentration in the aqueous solution stems from dissolved air. This transformation arises from the strong oxidization of H₂O₂ with a much higher electrode potential than that of H₂O [30]. In the presence of H₂O₂, Mg prefers to react with H₂O₂ without generating H₂ rather than to react with H₂O [54] and simultaneously Pt-catalyzed H₂O₂ decomposition can produce O₂ gas. This can be further verified by the fact that no bubble generates when bare Mg microparticles are put in H₂O₂ aqueous solution (Video S3).

As discussed above, the Mg-based micromotor can be driven by both Mg-H₂O reaction in the aqueous solution without H₂O₂ and Pt-catalyzed H₂O₂ decomposition when meeting H₂O₂. In the aqueous solution containing NaHCO₃ and PVP but without H₂O₂, the Mg core is consumed continuously via the Mg-H₂O reaction to produce numerous H₂ bubbles that propel the micromotor due to the conservation of momentum [55]. Figure 4 shows the Mg core amounts and motion speeds versus time, respectively. After overcoming the initial gravity-induced subsidence (Figure S4 and Video S4), the micromotor keeps moving by bubble recoils with the Mg core amount continuously decreasing as shown in Figure 4(a) and Video S5, indicating its self-consumed feature. At this time, the lifetime of the micromotor mainly depends on the opening size (Figure S5) and is in the range of a few minutes to dozens of minutes. With the Mg-H₂O reaction proceeding, there forms a confined space inside the micromotor, where the nucleation and growth of H₂ bubbles occur. At the early stages, the ejected bubbles are small, while they become large when the Mg core is almost gone. The size of the formed bubbles is mainly dependent on the growth time in the confined space. At the initial stages, the narrow space between the Mg core and the coating layer is too narrow for the growth of bubbles before they eject from the micromotor due to the capillary effect. Upon the enlarged confined space over time, the growth period is extended resulting in the formation of large bubbles. Figure 4(b) shows that the as-prepared micromotor has a relatively constant motion speed and consumption rate of the Mg core during Mg-H₂O reaction. This is not in agreement with the previous reported surface area-related reaction rate and can be explained by the effect of the confined space, in which the permeation of water is competitive with the escape of bubbles.

The as-constructed Mg-based micromotors can improve the motion controllability and lifetime by the spontaneous transformation of propulsion mechanisms from the self-consuming reaction to the catalytic reaction in terms of the surroundings. In H₂O₂ aqueous solution, the as-prepared Mg-based micromotor can be driven by Pt-catalyzed H₂O₂ decomposition and keeps constantly moving until the fuel (H₂O₂) runs out. The lifetime of the motor moving in H₂O₂ aqueous solution is determined by the amount of H₂O₂ and has been significantly prolonged because of the far excess “fuel” in the system. Figure 5 shows the effect of H₂O₂ concentration on the motion behaviors of the Mg-based micromotors. Taking a micromotor moving at 38°C for example, the as-prepared micromotor first accelerates and then decelerates the motion with increasing the H₂O₂ concentration (Figure 5(a)). This is dramatically different from the previously reported catalytic micromotors, whose speeds generally increase with increasing catalytic substrate concentrations [56–58]. The speeds of the Mg-based micromotor before and after the addition of H₂O₂ with different concentrations are shown in Figure S6. When the micromotor moves in the solution without H₂O₂, the motion speed is basically stable with an average value of about 13.2  μm/s depending on the relatively constant H₂ bubble size and generation frequency (Video S5). After adding H₂O₂, the motion speeds of the micromotor contributed by O₂ detachment from Pt-catalyzed H₂O₂ decomposition are closely related to H₂O₂ concentration. The initial increase of the speed with increasing H₂O₂ concentration (0~0.4 wt%) is attributed to an increased bubble generation frequency. The large variability of speed in 0.4 wt% H₂O₂ aqueous solution may be attributed to the rapid changes of \alpha and the bubble escape positions in motion (Video S6). The remarkable decrease of motion speed occurs as the H₂O₂ concentration is further increased to 0.8 wt%, implying a possible change of motion behavior. As shown in Figure 5(b) and Video S7, the micromotor displays a self-limitation motion like hovering due to the bilateral O₂ bubble generation and release as H₂O₂ concentration is higher than 0.8 wt%. This indirectly reduces the asymmetry of the micromotor and leads to the reduction of the speed. After diluting 0.8 wt% H₂O₂ aqueous solution to 0.67 wt%, the micromotor shows a curved ballistic trajectory as the O₂ bubbles are released from the opening instead of bilateral release, while the bilateral O₂ bubble generation and detachment occur again after supplementing H₂O₂ to 1.23 wt%. In other words, the as-prepared micromotor exhibits a self-limitation behavior in a relatively high-concentration H₂O₂ aqueous solution and returns to a curved ballistic motion at a low concentration. The transformation is reversible with changing the concentration of H₂O₂. Such unique concentration-induced self-limitation is associated with the permeability of H₂O₂ into the PNIPAM hydrogel layer. A high concentration of H₂O₂ aqueous solution can form a large concentration gradient between the two sides of the hydrogel layer to facilitate more H₂O₂ across the PNIPAM hydrogel layer and react with the middle Pt layer. As a result, the generated O₂ accumulates and reaches the heterogeneous nucleation energy on the side of the hydrogel layer for O₂ bubble nucleation and growth, resulting in a concentration-induced self-limitation.

To further illustrate the temperature/concentration-dependent self-limitation motion mechanisms of the Mg-based micromotor driven by Pt-catalyzed H₂O₂ decomposition, we have carried out numerical simulations over the Mg-based micromotor based on a 2D model (Figure S7) by chemical reaction engineering and dilute chemical species transport modules of COMSOL Multiphysics. Figure 6 shows the numerical simulation results of O₂ concentration distribution around the Mg-based micromotor with different temperatures and H₂O₂ concentrations. It should be noted that the simplified model only considers the O₂ concentration distribution for nucleation and ignores the fluid disturbance deriving from micromotor motion. In this model, the PNIPAM coating is treated as a porous media, because the swelling/shrinking of the hydrogel network can result in different porosities for O₂ diffusion, which is temperature-dependent. In a swelling state, a large cooperative diffusion coefficient benefits the permeation of H₂O₂ through the hydrogel layer resulting in a network with large porosity, which is approximately twice than that in a shrinking state [44, 59]. When the micromotor is moving below the LCST at a low-concentration H₂O₂ aqueous solution of 0.1 wt%, the O₂ generated inside the coating layer will reach the critical nucleation concentration of 68 mM within 1 ms due to the accumulative effect of the inner confined void (Figure 6(a)). Simultaneously, in view of the water-swelling feature of the PNIPAM hydrogel layer at this temperature, there are more H₂O₂ permeating across the hydrogel layer to react with the middle Pt layer resulting in O₂ nucleation on the side of the hydrogel layer. As a result, the micromotor shows a self-limitation behavior caused by bilateral O₂ bubble generation (Figure 2(a), iv). In contrast, by increasing temperature to 38°C, the access of H₂O₂ is slowed down by the contracted PNIPAM hydrogel layer. Consequently, the concentration of O₂ generated on the side of the hydrogel layer cannot reach its critical nucleation concentration under continuous outward diffusion (Figure 6(b)). Therefore, the inside generated O₂ bubble detachment from the opening is the origin of the sufficient propulsion of the as-prepared Mg-based micromotor (Figure 2(a), ii). However, increasing the H₂O₂ concentration to 0.8 wt% could facilitate the permeation of H₂O₂ aqueous solution into the PNIPAM hydrogel layer and the catalytic decomposition of H₂O₂ at the middle layer Pt (Figure 6(a)). This enables the O₂ generation on both sides of the middle Pt layer to reach the critical nucleation concentrations and thus induce self-limitation motion of the micromotor.

In this work, we present a kind of dual mechanism-propelled Mg-based micromotor with motion responsive to both temperature and H₂O₂ concentration. The as-constructed Mg-based micromotors are obtained by asymmetrically coating of Mg microspheres with catalytically active Pt and temperature-sensitive PNIPAM hydrogel layers in sequence. They can be propelled by Mg-H₂O reactions followed by Pt-catalyzed H₂O₂ decomposition when encountering H₂O₂ and exhibit extended motion lifetime. In the latter case, they would respond to both H₂O₂ concentration and temperature in solution environment, demonstrating unique temperature- and/or concentration-induced self-limitation motion like hovering. This is associated with the temperature-dependent diffusion coefficient and the concentration-related permeability of H₂O₂ aqueous solution in the PNIPAM hydrogel layer. Our results have for the first time demonstrated the Mg-based micromotors with motion responsive to external stimuli in a multifuel system and provided a strategy to push forward the development of intelligent Mg-based micromotors and to prolong the motion lifetime.