Self-Adaptive Flask-like Nanomotors Based on Fe3O4 Nanoparticles to a Physiological pH

In living bodies, pH values, which are precisely regulated and closely associated with diseased cells, can act as an efficient biologically intrinsic indicator for future intelligent biomedicine microsystems. In this work, we have developed flask-like carbonaceous nanomotors (FCNMs), via loading Fe3O4 nanoparticles (NPs) into a cavity, which exhibit a self-adaptive feature to a specific physiological pH by virtue of the pH-dependent dual enzyme-like activities of Fe3O4 NPs. Specifically, the peroxidase-like activity of Fe3O4 NPs in an acidic pH range, and the catalase-like activity in a near neutral and alkaline pH range, determine the products in the motion system (•OH, ions and O2), whose diffusions from the inner to the outside of the flask result in fluid movement providing the driving force for the movement of the FCNMs. Correspondingly, changes of the product concentrations and species in the physiological pH range (4.4–7.4) result, firstly, in velocity decrease and, then, with increase in pH, increase of the FCNMs occurs. Thanks to the non-linear velocity responsiveness, the FCNMs show intriguing pH taxis towards 6.8 (generally corresponding to the physiological pH in tumor microenvironments), where a maximum velocity appears. Furthermore, the superparamagnetic feature of the Fe3O4 NPs simultaneously endows the FCNMs with the abilities to be magnetic-oriented and easily separated. This work could significantly increase the possibility of nanomotors for targeted therapy of tumors and next-generation biotechnological applications.

Smart drug delivery with active targeting abilities is essential for futuristic biomedical fields. Compared with external field-propelled MNMs, chemically driven MNMs can harvest energy from their surroundings to induce movement, which provides a possibility for responding to local environments. In recent years, a growing effort has been devoted to engineer diverse chemically driven MNMs to respond to various stimuli, including metabolite concentration [12,13], microenvironment temperature [14,15], etc. In living bodies, pH value is an efficient biologically intrinsic indicator because it is highly diverse and precisely regulated in different organs and tissues. For example, the extracellular pH for healthy human tissues varies around 7.4, whereas the pH in tumor microenvironments is marginally acidic, ranging from 6.5 to 7.0 (average~6. 8), which further drops to 4.4-6.0 within tumor cells [16]. Minor changes of pH values in specific areas may affect physiological metabolism and even lead to diseases. Therefore, the various pH values can act as chemical signals to induce the responsive motion of MNMs. For example, the pH-responsive MNMs fabricated by catalytic materials, such as bio-enzymes [17], nanoenzymes [18][19][20][21] or active metals [22][23][24][25][26], the catalytic activities of which are influenced by environmental pH, usually show change of velocities and directions upon regulating the "fast/slow" aspect of pH-dependent driving reactions. Introducing pH-responsive polymers into MNMs, as gatekeepers to regulate the accessibility of "fuel", also easily achieve "on/off" of MNMs [21,27]. Additionally, pH-induced hydrophilicity/hydrophobicity transition and surface tension variation are also used to control the motion of MNMs displaying pH-responsiveness [28,29]. Although some well-designed MNMs can display a responsive motion behavior using pH as an indicator, the responsiveness levels are only indicated by monotonous velocity changes or transient on/off. However, besides simply identifying their microenvironments, the self-adaptive movement of MNMs to a specific pH, like some microorganisms with chemotaxis that are able to autonomously move to specific areas, is indispensable for futuristically smart motor systems. It means that they are able to seek focal sites in terms of a specific pH actively, and thus realize targeted drug delivery toward tumors.
Herein, we designed and constructed flask-like carbonaceous nanomotors (FCNMs) that exhibit self-adaptive pH-responsive motion behavior in the physiological pH range (4.4-7.4) on the basis of pH-dependent dual enzyme-like activities of Fe 3 O 4 NPs. The non-monotonic change of velocities for as-prepared FCNMs are influenced by the diffusion behaviors of pH-dependent products. At a lower pH range (4.4-5.6), Fe 3 O 4 NPs show peroxidase-like activity and the FCNMs are propelled by the diffusion of the product •OH and ions from the buffer solution used to balance the pH value of the system. With increase of the pH values, the peroxidase-like activity decreases until it disappears, while the catalase-like activity of Fe 3 O 4 NPs appears and increases. This leads to the transformation of the driving mechanisms from ionic and non-ionic diffusiophoresis to non-ionic diffusiophoresis. Correspondingly, the as-prepared FCNMs show a maximum motion velocity of 12.4 body lengths s −1 at pH 6.8 within 20 mM buffer solution and a fantastically spontaneous pH-chemotaxis from 7.4 to 6.8, which has a tempting application prospect for targeting tumor microenvironments. Additionally, thanks to the superparamagnetism of Fe 3 O 4 NPs, long-range motion control of the FCNMs is easily achieved by a magnetic field. These findings provide some insights into the pH-responsive motion mechanisms of Fe 3 O 4 -based MNMs, and offer some possibilities for application in biomedical fields, especially in the targeted therapy of tumors.

Synthesis of Fe 3 O 4 @OA NPs
The Fe 3 O 4 @OA NPs were synthesized according to previously reported procedure, with a slight modification [30]. Briefly, 3.6 g of Fe(oleate) 3 precursor, 1.14 g of oleic acid, and 3.22 g of oleyl alcohol were dissolved in 20 g of diphenyl ether at room temperature. The mixture was first heated to 90 • C at a constant heating rate of 8 • C/min in a nitrogen atmosphere, kept for 2 h to remove dissolved oxygen, and then heated to 220 • C to maintain this temperature for 10 min. After the reaction, the mixed solution was quickly cooled to room temperature, and the nanoparticles were washed three times with a mixed solvent of acetone and n-hexane by centrifugal separation, and finally dispersed in chloroform.

Modification of Fe 3 O 4 @OA NPs by DMSA
Fe 3 O 4 @DMSA NPs was prepared according to the previously reported method [31]. Briefly, 100 mg Fe 3 O 4 @OA NPs were first dispersed in 10 mL chloroform and 50 µL triethylamine were added. Next, 50 mg DMSA was dissolved in 10 mL DMSO and added to the above solution. The resulting mixed solution was magnetically stirred vigorously for 12 h at 60 • C. Then, the nanoparticles were washed three times with ethanol by centrifugal separation. In order to modify more DMSA on the surface of the Fe 3 O 4 NPs, and improve the water solubility, the nanoparticles were once again dispersed in ethanol for surface double-exchange. After repeating the previous operation, the obtained Fe 3 O 4 @DMSA NPs were washed with deionized water and dispersed in it.

Synthesis of Flask-like Carbonaceous Carriers (FCCs)
FCCs were prepared according to the previously reported method [32,33]. Briefly, 0.0365 g of sodium oleate and 0.0435 g of P123 were dissolved in 20 mL deionized water and stirred slowly for a period of time at 40 • C until it became clarified. Next, 40 mL aqueous solution, containing 3 g D-ribose, was added to the above solution. The mixed solution was transferred to a 100 mL autoclave and hydrothermally treated at 160 • C for 12 h. Then, the resulting FCCs were washed with deionized water three times.

Preparation of FCNMs
A certain number of dry FCCs were added to the water/ethanol mixed solution containing Fe 3 O 4 @DMSA NPs. Next, the mixture was treated in an ultrasonic instrument for 20 min to allow the FCCs to be fully dispersed and the Fe 3 O 4 @DMSA NPs to enter the hollow structure. The excess Fe 3 O 4 @DMSA NPs were removed by repeated washing and centrifugation until the supernatant was transparent. Then, the obtained products were placed in a vacuum drying oven at 60 • C for 12 h. The procedure was repeated three times and, finally, the resulting FCNMs were scattered in the deionized water.

Observation of the Motion of the FCNMs
The motion of the as-prepared FCNMs was observed by using a Leica DM 3000B optical microscope with a high-resolution CCD digital camera at a frame rate of 18 fps. For exploring the influence of pH values on motion behaviors, the H 2 O 2 concentration of the whole system was diluted to 2.5 wt%, and their pH values were adjusted to 4.4, 5.0, 5.6, 6.2, 6.8 and 7.4 using 20 mM disodium hydrogen phosphate-citrate buffer solution. For studying the influence of H 2 O 2 concentrations on motion behaviors, concentrations of 0.5 wt%, 1 wt%, 2.5 wt%, 5 wt% and 10 wt% were adopted, respectively. For the spontaneous chemotactic experiments, agar (0.5 cm × 0.5 cm) was put in 2.5 wt% H 2 O 2 solution (pH 6.8) for 12 h at room temperature in advance, then taken out and put on a hydrophobic glass substrate. Subsequently, 2.5 wt% H 2 O 2 solution (pH 7.4), containing the FCNMs, was dropped on the other side. For the magnetic manipulation experiments for the FCNMs, 2.5 wt% H 2 O 2 solution (pH 6.8), containing the FCNMs, was dropped on a hydrophobic glass substrate, and the direction of its motion was manipulated by moving an external magnet.

Characterization
The scanning electron microscope (SEM) images were taken with a Hitachi S-4800 microscope (Tokyo, Japan). The transmission electron microscopy (TEM) and energydispersive X-ray (EDX) images were taken with a FEI F200 microscopy (Hillsboro, OR, USA). The powder X-ray diffraction (XRD) patterns were obtained with a Empyrean diffractometer (Almelo, Holland). Magnetic properties were analyzed with a Lake Shore 7400 vibrating sample magnetometer (VSM, Westerville, OH, USA). The thermal properties were evaluated with a STA 449F3 thermogravimeter (TG, Selb, Germany). The dynamic light scattering (DLS) measurements were proceeded with a NanoBrook 90Plus Zeta (New York, NY, USA). The motion videos and images were recorded with a Leica DM 3000B optical microscope (Wetzlar, Germany).

Numerical Simulation
Our numerical model was implemented in COMSOL Multiphysics (version 5.4a) in a 2D axisymmetric configuration to reduce the computational cost. The primary model parameters are shown in Table S1 and Figure S1 in the Supplementary Materials. The chemical reaction engineering and dilute chemical species transport module were employed. Fe 3 O 4 NPs were presumed to distribute uniformly in the flask bottom with a constant concentration of 7.14 × 10 −5 mol·m −2 . When the FCNMs moved in an acid condition (pH 4.4), chemical reactions: in the confined cavity, which were calculated by a chemical reaction engineering model with a rate constant of 65 M −1 ·s −1 and 1.1 × 10 10 M −1 ·s −1 , respectively, and the governing equation was as follows: Here, for a species i, c i is the concentration, R i is the rate of the species, R ads,i is the surface reaction area, and V r is the reactor volume. Transport of diluted species is used to compute the concentration field with the governing equation: where J i and c i are the ionic flux and ionic concentration for species i, respectively, u is the flow velocity, D is the ionic diffusion coefficient. We noted that diluted species distributions were affected by flow field and chemical potential field. Chemical reaction: Rather, the reaction: place on the catalytic surface, and the boundary fluxes at the catalytic surfaces thus became: The FCNM was driven by both chemical potential fields generated by concentrations of ionic species and non-ionic species, respectively. As the FCNM was fixed in the model, we set the FCNM surface as the slip boundary condition, and combined chemiosmotic flow, with a slip velocity, as: where b p_ionic and b p_non−ionic are the surface mobility of the FCNM, caused by concentrations of ionic species and non-ionic species. On account of solvation of ions, b p_ionic and b p_non−ionic have opposite signs, which means opposite direction of flow. And the quantity (I − nn) defines the concentration gradients, where I denotes the second-order unit tensor. Fluid field around the FCNM was solved by Stokes equation: where p is the pressure, η is dynamic viscosity of water. When the FCNMs moved in a near alkaline condition (pH 7.4), the chemical potential field caused by ∇c O 2 was considered the main source of driving force. The boundary conditions at the outer edge of the simulation domain were chosen to represent the bulk: constant concentrations of species, and a no stress boundary condition for fluid. At the FCNMs' surfaces, we prescribed a uniform charge density. We solved for transient state for the FCNMs, which shows the system reached a stable state after 1 ms, hence, results in the discussion all are at 1 ms.

Implementation Strategy of the pH-Responsive FCNMs
The self-adaptive physiological pH-responsive FCNMs were realized by virtue of the dual-enzyme catalytic reactions of Fe 3 O 4 NPs for H 2 O 2 decomposition inside the confined cavity of flask-like structures ( Figure 1). When Fe 3 O 4 NPs encountered the catalyzed substrate of H 2 O 2 , the following reactions occurred: rials 2021, 11, x Figure 1. Schematic of the self-adaptive pH-responsive motion behaviors of the as-prepared based on the pH-dependent catalytic activities of Fe3O4 NPs. The velocity and driving force the flask mouth direction were set to be positive (V+ and F+). At a lower pH value, the FCNM driven by the co-effect of C6H7O7-and •OH diffusiophoresis in the opposite directions. W reduction of ionic diffusiophoresis and the addition of O2 diffusiophoresis, the motion increased in the weak acid environment. In the near alkaline range, O2 diffusiophoresis p the driving force for the FCNMs showing a reduction of the velocity.

Preparation and Characterization of the FCNMs
The FCNMs were prepared by loading Fe3O4@DMSA NPs in flask-like carbon carriers (FCCs), as schematically illustrated in Figure 2a. The FCCs were synthesiz soft-template-assisted hydrothermal method with a relatively uniform size of 60 cavity of ~280 nm and a flask mouth of ~140 nm ( Figure S2). Their surfaces were hydroxyl and carboxyl groups giving rise to a lower surface potential (−57 mV) At a lower pH (e.g., pH < 5.6 in Figure 1), Equation (6) dominated with a rate constant of 65 M −1 ·s −1 to produce •OH, and Equations (7) and (8) were inhibited because of the acidity of the solution, indicating its peroxidase-like activity. To eliminate the effect of ion by-products on the pH value, all motions of the as-prepared FCNMs were observed in a 20 mM buffer solution. At this time, the FCNMs were driven by the co-effect of C 6 H 7 O 7 − and •OH diffusiophoresis in the opposite directions. With increasing pH to near neutral or alkaline, Equation (7) was the rate-determining step (k = 2.5 × 10 −3 M −1 ·s −1 ) was promoted, and O 2 was produced instead of •OH (Equation (8)). This could lead to the driving force having to convert into O 2 diffusiopheresis. Therefore, with increasing pH in the acidic environment (pH < 5.6), the decrease of peroxidase-like activity could reduce the concentrations of C 6 H 7 O 7 − and •OH, and, thus, decrease the motion velocity of the FCNM. On further increasing pH from 5.6 to 6.8, the diffusiophoresis of C 6 H 7 O 7 − was weakened until it disappeared, because the occurrence of Equations (7) and (8) (7), the concentration of O 2 was lower in the system and the velocity decrease might have also been due to the diffusion restriction of the flask-like structures. As a result, the as-prepared FCNMs displayed self-adaptive motion behaviors following the pH changes in the system.

Preparation and Characterization of the FCNMs
The FCNMs were prepared by loading Fe 3 O 4 @DMSA NPs in flask-like carbonaceous carriers (FCCs), as schematically illustrated in Figure 2a. The FCCs were synthesized by a soft-template-assisted hydrothermal method with a relatively uniform size of 600 nm, a cavity of~280 nm and a flask mouth of~140 nm ( Figure S2). Their surfaces were rich in hydroxyl and carboxyl groups giving rise to a lower surface potential (−57 mV). Then, using the thermal decomposition method, different Fe 3 O 4 @OA NPs were synthesized by regulating experimental parameters, including aging temperature ( Figure S3), aging time ( Figure S4) and raw material molar ratios ( Figure S5). The results indicated that the as-synthesized nanoparticles were magnetite (Fe 3 O 4 ; JCPDS no. 75-0033) with a cubic inverse spinel structure, and simultaneously, it could also be found that the particle size and saturation magnetization increased with increasing aging temperature, time or decreasing the amount of oleyl alcohol. Subsequently, the nanoparticles with the highest catalytic activity ( Figure S6 and Table S2) were modified by DMSA to achieve transformation from hydrophobicity to hydrophilicity ( Figure S7). Next, the morphology and structure of the as-prepared FCNMs were characterized by SEM ( Figure 2b

Motion Behaviors of the FCNMs at Different pH Values
By virtue of the dual enzyme-like activities of Fe3O4 NPs that can react with H2O2 to produce •OH in acidic environment (peroxidase-like activity) and O2 in near neutral and alkaline environments (catalase-like activity), it is conceivable that the as-prepared FCNMs could exhibit pH-dependent motion behaviors. In term of this, we explored the motion behaviors of the FCNMs at 2.5 wt% H2O2 aqueous solution in physiological pH values ranging from 4.4 to 7.4 with a gradient of 0.6. Since H2O2 decomposition usually causes the pH change in solution, the motion behaviors were conducted in a 20 mM buffe solution and observed using an inverted optical microscope. In each case, at least 15 FCNMs were analyzed. Typical tracking trajectories of the FCNMs with, or without, H2O "fuel" are presented in Figure 3a. In the absence of H2O2, the FCNMs performed Brownian motion only (Video S1), while those with the presence of H2O2 exhibited active self propulsion, indicating that the driving force for the motion derived from the reaction between FCNMs and H2O2. The motion behaviors of the FCNMs exhibited a non-linear change with increasing pH values, which was further verified by the mean-square displacement (MSD) versus the time interval (Δt) curves, according to the extracted 2D coordinates along the trajectories (Figure 3b). Without adding H2O2, the linear relationship between the MSD and Δt indicated a typical Brownian diffusive motion. The parabolic MSD curves, when using H2O2 as the fuel, suggested directional self-propulsion of the FCNMs. In order to clarity the relationship, we calculated the average velocities ( Figure  3c) and diffusion coefficients (Figure 3d) of the FCNMs versus the pH values, which suggested their self-adaptive pH responsiveness. The velocities of the FCNMs decreased with increasing pH values from 4.4 to 5.6, and then rose until pH value was up to 6.8 with a velocity of 7.42 ± 1.0 μm/s (≈12.4 body lengths·s −1 ), after which the velocities subsequently dropped when pH was 7.4 (Videos S2 and S3). A similar dependence could

Motion Behaviors of the FCNMs at Different pH Values
By virtue of the dual enzyme-like activities of Fe 3 O 4 NPs that can react with H 2 O 2 to produce •OH in acidic environment (peroxidase-like activity) and O 2 in near neutral and alkaline environments (catalase-like activity), it is conceivable that the as-prepared FCNMs could exhibit pH-dependent motion behaviors. In term of this, we explored the motion behaviors of the FCNMs at 2.5 wt% H 2 O 2 aqueous solution in physiological pH values ranging from 4.4 to 7.4 with a gradient of 0.6. Since H 2 O 2 decomposition usually causes the pH change in solution, the motion behaviors were conducted in a 20 mM buffer solution and observed using an inverted optical microscope. In each case, at least 15 FCNMs were analyzed. Typical tracking trajectories of the FCNMs with, or without, H 2 O 2 "fuel" are presented in Figure 3a. In the absence of H 2 O 2 , the FCNMs performed Brownian motion only (Video S1), while those with the presence of H 2 O 2 exhibited active self-propulsion, indicating that the driving force for the motion derived from the reaction between FCNMs and H 2 O 2 . The motion behaviors of the FCNMs exhibited a non-linear change with increasing pH values, which was further verified by the mean-square-displacement (MSD) versus the time interval (∆t) curves, according to the extracted 2D coordinates along the trajectories (Figure 3b). Without adding H 2 O 2 , the linear relationship between the MSD and ∆t indicated a typical Brownian diffusive motion. The parabolic MSD curves, when using H 2 O 2 as the fuel, suggested directional self-propulsion of the FCNMs. In order to clarity the relationship, we calculated the average velocities ( Figure 3c) and diffusion coefficients (Figure 3d) of the FCNMs versus the pH values, which suggested their self-adaptive pH responsiveness. The velocities of the FCNMs decreased with increasing pH values from 4.4 to 5.6, and then rose until pH value was up to 6.8 with a velocity of 7.42 ± 1.0 µm/s (≈12.4 body lengths·s −1 ), after which the velocities subsequently dropped when pH was 7.4 (Videos S2 and S3). A similar dependence could also be found in diffusion coefficients of the FCNMs at different pHs calculated by the equation of MSD = 4D eff ∆t. This confirmed that the as-prepared FCNMs could self-regulate their velocities, by sensing pH values in the physiological range, which provided a possibility for their self-adaptive movement toward a specific pH.

Experimental Verification and Simulation Analysis for the pH-Responsive Mechanisms
In order to figure out why the FCNMs could exhibit pH-responsive motion behavior, we conducted a deep exploration for propulsion mechanisms through experimental verification and simulation analysis. The relationship between the dual enzyme-like activities of Fe3O4 NPs and environmental pH was studied firstly. The peroxidase-like activity of Fe3O4 NPs was tested by catalyzing the oxidation of substrate 3,3′,5,5′tetramethylbenzidine (TMB) in the presence of H2O2, which was accompanied by color change to blue and appeared as obvious UV-Vis absorption at 650 nm ( Figure S8). The results showed that the peroxidase-like activity increased with increasing H2O2 concentrations and reached the maximum at 2.5 wt%, whereas it decreased with increasing the pH values till at pH 7.4 there was no •OH production. Instead, when pH > 5.6, O2 generation could be detected by monitoring dissolved oxygen in the reaction system, indicating the catalase-like activity of the as-synthesized Fe3O4 NPs. Then, we focused on verifying whether this law held true once the Fe3O4 NPs were encapsulated in a confined space to form a motor. The changes of the light absorption value at 650 nm in the first 1000 s of the reaction system were plotted in Figure 4a. It could be found that as the pH increased, the slope values in the linear range decreased and became zero once the pH exceeded 6.8, and, at the same time, the amounts of O2 generation increased ( Figure  4b). This suggested that, with increase of the pH values, Equations (7) and (8) could be activated to expend •OH, resulting in the decrease until it disappeared in the system. Further increasing the pH values, O2, instead of •OH, became the main product. Therefore, in terms of the results it was concluded that with increasing pH values from 4.4 to 7.4, the main products of the system experienced three stages: only •OH generated, •OH and O2

Experimental Verification and Simulation Analysis for the pH-Responsive Mechanisms
In order to figure out why the FCNMs could exhibit pH-responsive motion behavior, we conducted a deep exploration for propulsion mechanisms through experimental verification and simulation analysis. The relationship between the dual enzyme-like activities of Fe 3 O 4 NPs and environmental pH was studied firstly. The peroxidase-like activity of Fe 3 O 4 NPs was tested by catalyzing the oxidation of substrate 3,3 ,5,5 -tetramethylbenzidine (TMB) in the presence of H 2 O 2 , which was accompanied by color change to blue and appeared as obvious UV-Vis absorption at 650 nm ( Figure S8). The results showed that the peroxidase-like activity increased with increasing H 2 O 2 concentrations and reached the maximum at 2.5 wt%, whereas it decreased with increasing the pH values till at pH 7.4 there was no •OH production. Instead, when pH > 5.6, O 2 generation could be detected by monitoring dissolved oxygen in the reaction system, indicating the catalase-like activity of the as-synthesized Fe 3 O 4 NPs. Then, we focused on verifying whether this law held true once the Fe 3 O 4 NPs were encapsulated in a confined space to form a motor. The changes of the light absorption value at 650 nm in the first 1000 s of the reaction system were plotted in Figure 4a. It could be found that as the pH increased, the slope values in the linear range decreased and became zero once the pH exceeded 6.8, and, at the same time, the amounts of O 2 generation increased (Figure 4b). This suggested that, with increase of the pH values, Equations (7) and (8) Figure 5 presents the effect of H2O2 concentrations on the motion behaviors of the asprepared FCNMs. In each case, at least 15 FCNMs were analyzed. The results display that the FCNMs at 0 wt% H2O2 performed Brownian motion only, while those with the presence of H2O2 exhibited active self-propulsion with various trajectories, dependent on the H2O2 concentrations (Figure 5a). Parabolic MSD curves in Figure 5b show increasing slope values in the linear range until the H2O2 concentration was up to 2.5 wt%, also suggesting a transition from Brownian diffusive motion to self-propulsion. We also calculated the average velocities and diffusion coefficients of the FCNMs versus the H2O2 concentrations. The mobility of the FCNMs continued to increase with increasing H2O2 concentrations at the beginning, and dropped obviously when the H2O2 concentration was higher than 2.5 wt% (Figures 5c,d). It is suggested that more H2O2 might increase the ionic strength in the system, which had a negative effect on the motor movement. This phenomenon is consistent with the description of Sen et al., where it was proposed that when the total ionic strength of the system increases, there is a great negative effect on the motion capability of the motors driven by ionic diffusiophoresis [34]. The typical video of the FCNM swimming in 2.5 wt% H2O2 aqueous solution is provided in Video S5 with a maximum average velocity of 7.01 ± 0.59 μm/s (≈11.7 body lengths·s −1 ). To further elucidate the self-adaptive pH-responsive motion behaviors of the FCNMs, we simulated the flow field distributions at different pH values, respectively, using COM-SOL multiphysics based on a 2D model. On the basis of Figure 4a,b, only Equation (6) occurred at a lower pH (e.g., pH < 5.6), •OH was the main product accompanied by the generation of OH − ( Figure S9). In order to exclude the effect of OH − on the pH value, Equation (9) occurred and the ionic product C 6 H 7 O 7 − was generated. Therefore, the FC-NMs in this case were co-driven by ionic and non-ionic diffusiophoresis. As shown in Figure 4c, the ionic concentration gradient formed at the opening, while the non-ionic concentration gradient was at the inner neck in an opposite direction. Therefore, the motion direction of the FCNMs depended on their relative magnitude. The simulated result showed that non-ionic diffusiophoresis played a dominant role in the current system, which drove the motor moving toward the flask mouth direction (Figure 4d). Limited by the resolution of the used optical microscope, the motion direction of the as-prepared FCNMs could not be distinctly resolved. Therefore, we treated much larger FCCs with a similar structure (~1 µm) to fabricate micromotors to experimentally verify the motion direction ( Figure S10 and Video S4), which was in good agreement with the simulated result, and suggested the dominant effect of non-ionic diffusiphoresis. The reduced motion velocity in acidic pH from 4.4 to 5.6 could be attributed to the reduction of peroxidase-like activity of Fe 3 O 4 NPs. When the pH value was up to 5.6, Equations (7) and (8) were activated to not only consume •OH groups but also to produce H + , which was evidenced by the appearance of dissolved O 2 in the system. At this time, OH − ions produced from Equation (6) were partly neutralized by H + ions giving rise to the reduction of ion concentrations, and thus weakening the ionic diffusiophoresis. Additionally, O 2 generation could, to some extent, increase the non-ionic dissusiophoresis, and, thus, accelerate the FCNMs. With the disappearance of ionic diffusiophoresis in the system and the complete conversion of •OH groups, in terms of Equation (8), the movement velocity of the FCNMs reached a maximum at pH 6.8. The subsequently reduction of the velocity in near alkaline range might derive from the relatively low O 2 concentration gradient ( Figure S11) and the diffusion restriction of the flask-like structures. From the point of view of the relative strength of the flow fields, we could also find that the non-ionic diffusiophoresis from •OH was much stronger than that of O 2 , which was in keeping with their concentration gradient distribution and the experimental motion behavior.  Figure 5b show increasing slope values in the linear range until the H 2 O 2 concentration was up to 2.5 wt%, also suggesting a transition from Brownian diffusive motion to self-propulsion. We also calculated the average velocities and diffusion coefficients of the FCNMs versus the H 2 O 2 concentrations. The mobility of the FCNMs continued to increase with increasing H 2 O 2 concentrations at the beginning, and dropped obviously when the H 2 O 2 concentration was higher than 2.5 wt% (Figure 5c,d). It is suggested that more H 2 O 2 might increase the ionic strength in the system, which had a negative effect on the motor movement. This phenomenon is consistent with the description of Sen et al., where it was proposed that when the total ionic strength of the system increases, there is a great negative effect on the motion capability of the motors driven by ionic diffusiophoresis [34]. The typical video of the FCNM swimming in 2.5 wt% H 2 O 2 aqueous solution is provided in Video S5 with a maximum average velocity of 7.01 ± 0.59 µm/s (≈11.7 body lengths·s −1 ).

Chemotactic Motion and Magnetic Responsiveness of the FCNMs
Thanks to the unique pH-responsive motion behaviors of FCNMs in the physiological range, they can display a self-adaptive chemotaxis motion toward a specific pH where they exhibit a maximum velocity. As shown in Figure 6a, a left-to-right gradually decreasing pH gradient was constructed by using a saturated agar filled with

Chemotactic Motion and Magnetic Responsiveness of the FCNMs
Thanks to the unique pH-responsive motion behaviors of FCNMs in the physiological range, they can display a self-adaptive chemotaxis motion toward a specific pH where they exhibit a maximum velocity. As shown in Figure 6a, a left-to-right gradually decreasing pH gradient was constructed by using a saturated agar filled with fuel solution at pH 6.8, which was placed on the other side of the mixed solution at pH 7.4 containing the FCNMs. In the presence of 2.5 wt% H 2 O 2 , the tracking trajectories of the 10 FCNMs were observed, and are depicted in Figure 6a. They all exhibited chemotactic behavior from pH 7.4 to 6.8 (Video S6). Moreover, acceleration phenomena in the process of getting closer to the agar were observed for the 10 FCNMs by calculating their real-time velocities (Figure 6b). The average velocities of the corresponding FCNMs in the first 10 s and the last 10 s were also calculated in Figure S12. To avoid any accidental effect, we cleverly set up an opposite pH gradient to observe the motion direction of the FCNMs positioned on the side of the mixed solution at pH 6.8. As expected, the FCNMs were far away from the saturated agar filled with fuel solution at pH 7.4 ( Figure S13 and Video S7). Therefore, the results provided a direction control strategy for the as-prepared FCNMs, which was capable of actively targeting a specific area with a unique pH value. Besides this, on-demand direction manipulation could also be achieved by external magnetic field because of the inherent superparamagnetism of Fe 3 O 4 NPs (Figure 6c). Facilely magnetic separation and easy re-dispersion features also endowed the as-prepared FCNMs with excellent recycling characters (Figure 6d).

Conclusions
In this work, we have demonstrated flask-like carbonaceous nanomotors (FCNMs) exhibiting self-adaptiveness toward a specifically physiological pH on the basis of the dual enzyme-like activities of magnetic Fe3O4 NPs. In the process of increasing pH in the physiological range from 4.4 to 7.4, the peroxidase-like activity for Fe3O4 NPs gradually wears off, accompanied by enhanced catalase-like activity, resulting in a transformation of the propulsion mechanisms from non-ionic and ionic diffusiophoresis to single non-

Conclusions
In this work, we have demonstrated flask-like carbonaceous nanomotors (FCNMs) exhibiting self-adaptiveness toward a specifically physiological pH on the basis of the dual enzyme-like activities of magnetic Fe 3 O 4 NPs. In the process of increasing pH in the physiological range from 4.4 to 7.4, the peroxidase-like activity for Fe 3 O 4 NPs gradually wears off, accompanied by enhanced catalase-like activity, resulting in a transformation of the propulsion mechanisms from non-ionic and ionic diffusiophoresis to single non-ionic diffusiophoresis, which ultimately feeds back to a non-linear responsiveness upon pH. This endows the FCNMs with a unique chemotaxis towards a slightly acidic region (pH 6.8) that is a specific pH for generally physiological focal sites, such as tumor microenvironments. These results provide a design strategy for MNMs exhibiting a self-adaptive physiological pH-responsive motion, and the characteristic pH taxis, as well as the magnetic control, also offer a larger possibility in future biotechnological applications, especially for actively targeted tumor therapy.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/nano12122049/s1, Figure S1: Meshing conditions of the theoretical model. Figure S2: Characterizations of the FCCs. Figure S3: Characterizations of the Fe 3 O 4 @OA NPs synthesized at different aging temperatures. Figure S4: Characterizations of the Fe 3 O 4 @OA NPs synthesized at different aging time. Figure S5: Characterizations of the Fe 3 O 4 @OA NPs synthesized at different raw material molar ratios between oleyl alcohol and oleic acid. Figure S6: Relative peroxidase-like activities of the as-synthesized Fe 3 O 4 @OA NPs calculated by taking the group with the highest activity as 100%. Figure S7: Characterizations of oil-soluble Fe 3 O 4 @OA and water-soluble Fe 3 O 4 @DMSA NPs. Figure S8: Verifications of dual enzyme-like activities of Fe 3 O 4 @DMSA NPs. Figure S9: Verification of OHproduct by monitoring the change of pH values of pure 2.5 wt% H 2 O 2 aqueous solution, FCNMs aqueous solution and their mixed solutions in 30 min. Figure S10: Timelapse images of a comparable FCNM with a larger size of~1 µm at pH 4.4. Figure S11: Concentration gradient distributions of •OH, C 6 H 7 O 7 − and O 2 on the symmetric axis of the FCNMs. Figure S12: A discrete distribution of the average velocities of different chemotactic FCNMs in the first 10 s and last 10 s. Figure S13: Chemotactic track trajectories of the FCNMs in a right-to-left gradually decreasing pH gradient at 2.5 wt% H 2 O 2 in 30 s. Table S1: Typical parameters used in the simulations. Author Contributions: Investigation, T.G.; software, J.L.; data curation and writing-original draft preparation, T.G. and L.X.; writing-review and editing, T.G., L.X. and J.G.; conceptualization and supervision, L.X. and J.G. All authors have read and agreed to the published version of the manuscript.
Funding: This research received no external funding.