1 Introduction

The 2023 European Drug Report^[1] issued an alert about the great diversity, availability, and use of drugs of abuse including illicit drugs and misused prescription medicines. Unfortunately, when considering the impact that these psychoactive substances are generating, not only health risks but also an environmental threat must be discussed. They easily penetrate the hydrological cycle as unaltered forms or metabolites excreted after human consumption or as residues from production/distribution activities from the drug market. Hence, drugs of abuse have become an emergent class of pollutants that endangers freshwaters and marine habitats.^[2] Among them, special attention must be devoted to methamphetamine (MA). The growing consumption of this highly addictive element, as well as its use as a medical treatment for some mental health diseases,^[3] has led to the appearance of an increasing number of clandestine manufacturers as well. Despite official guidelines on MA decontamination,^{[4, 5]} the inefficiency of current remediation methods^{[6, 7]} has caused the detection of increasing MA concentrations in wastewater analyses in up to 40 European cities in the 2021–2022 period.^[1] Besides, a harmful effect on aquatic ecosystems has been proved by the behavior alteration of fish.^[8] In addition, MA residues have been detected in tap water in several European countries,^[6] turning MA pollution into an additional public health issue. Consequently, it is necessary to test alternative cleaning methods to remediate this situation.

Microrobots enable enhanced mass transfer due to their autonomous motion, great tunability of their surfaces for the development of specific tasks,^[9-14] and their swarming/collective behavior. Consequently, they are rapidly evolving in their applications for environmental remediation.^[15] In fact, their action against different residues, from organic dyes^{[16, 17]} to microplastics,^[18-21] heavy metals,^[22-24] hormones,^[25] or antibiotics^{[26, 27]} among others, have been discussed. However, the utilization of microrobots for the removal of drugs of abuse remains largely unexplored.^[28]

Here, we utilize magnetically powered microrobots modified with chiral cyclodextrin (CD) molecules.^{[29, 30]} CDs are organic supramolecular structures with hydrophobic cavities that provide an optimal environment to form host–guest complexes based on non-covalent interactions. CDs are often used for drug delivery and encapsulation and have also been extensively used for the removal of different water pollutants^[31-33] including MA.^[34] In fact, the host–guest complex between CDs and MA has been widely reported in the literature, particularly in analytical chemistry, where CD-based stationary phases of chromatographic techniques are used for the MA separation and subsequent detection.^[35-39] However, they are scarcely employed in the field of microrobots, being limited to a few proof-of-concept studies performed on enantioselectivity^[40] and drug delivery.^{[41, 42]}

This work puts together promising microrobot technology and CD chemistry to deal with the environmental and health threats caused by MA pollution. To this end, hybrid urchin-like hematite (α-Fe₂O₃) microparticles carrying magnetite (Fe₃O₄) nanoparticles are combined with CD surface functionalization for MA removal. The iron oxide-based microrobots provide magnetic maneuvering that benefits enhanced mass transfer while CD hydrophobic cavities promote MA capture (Scheme 1).

2 Results and Discussion

Hematite microparticles, the most stable iron oxide derivative, were used as backbone material for mag-CD microrobots. To date, hematite-based microrobots showing spheres,^[43] cubes,^[44] dendrites,^[45] or peanut-like^[46] morphologies have already been explored. The hydrothermal method selected in this work provided original urchin-like shapes with an approximate diameter of 3.6 ± 0.9 µm (Figure 1A) and a homogeneous distribution of iron and oxygen, their constituent elements (Figure 1B). As depicted in Figure 1C (top), the XRD analysis confirmed a perfect indexation to the α-Fe₂O₃ rhombohedral crystal system (100%, ICDD: 01-076-4579), where all peaks were easily identified. The Brunauer–Emmett–Teller (BET) method was used to measure a specific surface area of 19.3 m² g⁻¹ (Figure S1, Supporting Information). This value is slightly higher than the one reported for hollow urchin-like hematite particles (18.8 m² g⁻¹),^[47] which is beneficial for their further surface functionalization.

β-CD molecules (Figure 1C (central)) were anchored to pristine hematite urchins through an amine modification in line with previous literature^[48] (Figure S2, Supporting Information, see Experimental Section for further details). FTIR spectra were recorded to gain more insights into the presence of these organic cavities (Figure 1D). In addition to the Fe─O stretching modes detected below 550 cm⁻¹,^{[49, 50]} new features were observed in the case of the modified urchins and associated with β-CDs: O─H stretching, and C─H asymmetric/symmetric stretching modes were identified at 3300 and 2921 cm⁻¹, respectively, while the H─O─H bending was found at 1643 cm⁻¹. On the other hand, C─O─C asymmetric and symmetric stretching vibrations were associated with the bands at 1151 and 1022 cm⁻¹, respectively.^[51-53]

It is known that α-Fe₂O₃ hematite-based microrobots respond to relatively weak magnetic fields.^{[45, 54]} Nonetheless, different chemical/physical factors, such as size, morphology, and vacancies, can alter the magnetic coercivity in α-Fe₂O₃ microparticles.^[55-57] The magnetic properties of the urchin-like hematite microparticles (α-Fe₂O₃) and α-Fe₂O₃@CD were analyzed by means of a vibrating sample magnetometer (VSM); however, only insignificant magnetic hysteresis was observed, indicating their low ability to be magnetized (Figure S4, Supporting Information). The surface of the microurchins was modified with commercially available Fe₃O₄ NPs²¹ (Figure S3, Supporting Information) to obtain mag-CD microrobots (α-Fe₂O₃@CD/Fe₃O₄). Ferromagnetic behavior was demonstrated by the resulting magnetic hysteresis loop as measured by VSM due to the presence of the Fe₃O₄ NPs (Figure S4, Supporting Information). Moreover, a positive response was observed when exposed to a permanent magnet (Figure S5, Supporting Information), corroborating once more the magnetic character of mag-CD microrobots. The whole preparation method was monitored by zeta potential measurements (Figure 1E), showing a clear evolution from the pristine urchins to final mag-CD microrobots as follows: −10 ± 2 mV (Fe₂O₃); 25.4 ± 0.3 mV (Fe₂O₃@CD); 15 ± 1 mV (Fe₂O₃@CD/Fe₃O₄). Additionally, Figure 1C (bottom) includes XRD data for mag-CD microrobots, showing all main contributions from Fe₃O₄ NPs (ICDD: 01-088-0866), as well as the features at 2θ < 20° that could be associated with β-CDs^[51] (see Figure S6, Supporting Information for further details). It is also worth mentioning that the original urchin-like morphology was preserved after the functionalization process as demonstrated by the scanning electron microscope (SEM) images of a mag-CD microrobot in Figure 1F.

The magnetic maneuvering of hybrid mag-CD microrobots under a transversal rotating magnetic field was faced after concluding an exhaustive material characterization. Figures 2 and S7, Supporting Information, and Movies S1–S3, Supporting Information, summarize the most relevant findings on this matter. First, the navigation of a single microrobot under a predefined linear mode was checked using a fixed field intensity of 5 mT and several frequencies (1–5 Hz). Figure 2A (Movie S1, Supporting Information) depicts the clear net displacements recorded in all cases, showing how microrobots covered greater distances with increasing frequency.

Alternative predefined modes, such as perpendicular (Figure 2B and Movie S2 (left), Supporting Information) or random (Figure 2C and Movie S2 (right), Supporting Information), were also explored, proving total spatial control over their navigation. Importantly, different-sized microrobots were observed during the analysis of the different trajectories. The smallest entities were identified as residual Fe₃O₄ NPs while bigger aggregates of mag-CD microrobots were thought to have been formed due to the electrostatic interactions occurring during their preparation and/or navigation.^[20] Nonetheless, this size variability did not have any impact on the magnetic maneuvering. On the contrary, harmonized motions following a swarming effect were recorded in all cases. Velocity data were also considered during the motion analysis of mag-CD microrobots and estimated at different frequencies (Movie S3 and Figure S7A, Supporting Information). As expected, higher frequencies resulted in faster navigation, in line with the direct proportionality between both terms up to 6 Hz. Beyond this frequency, the velocity of mag-CD microrobots became unstable and decreased. Therefore, the out-step frequency of the mag-CD microrobots was estimated at 6 Hz. In addition, the performance of the microrobots was evaluated in alternative realistic aqueous media (tap and river water). The microrobots could move freely in both samples (Movie S3, Supporting Information) although some differences in their speed trends were noticed. As depicted in Figure S7B, Supporting Information, faster navigation in tap water was estimated when increasing frequency from 1 to 6 Hz was applied, and similarly in DI water. Nonetheless, a different development was found in river water when no apparent evolution was observed beyond 2 Hz. This finding may be related to the effect that alternative aqueous matrixes and their viscosities have on the final performance of the microrobots, shifting their out-step frequency.

Finally, the removal of illicit drug residues by mag-CD microrobots was addressed, taking MA as the target pollutant (Figure 3A). A magnetic assay named “dynamic” was carried out with a dispersion of mag-CD microrobots (0.8 mg mL⁻¹) in a 200 ng mL⁻¹ MA aqueous solution without further additives. A transversal magnetic rotating field (5 mT, 3 Hz) was applied in this “dynamic assay” for 40 min with a predefined random propulsion mode.^[58] In addition, high-resolution mass spectroscopy (HRMS) was used to evaluate the fate of MA. Considering a positive atmospheric pressure chemical ionization (APCI) method, the presence of MA was associated with a signal observed at m/z = 150.127 ([C₁₀H₁₅N+H]⁺), and the evolution of this feature was related to the action of the microrobots. Interestingly, the original MA signal (Figure 3B (top)) was considerably reduced after the performance of mag-CD microrobots (Figure 3B (bottom)), achieving a significant %MA removed: 50.5 ± 15.4. It is known that basic pH values like 10–11 enhance the kinetics of the CD-MA host–guest complexes.^{[34, 59]} Nonetheless, the performance of the mag-CD microrobots was analyzed without optimizing this parameter, in line with the actual environments of raw waters where the micromotors would be tested as microcleaners. It is worth emphasizing that the presence of MA in surface water, potable water, and wastewater is detected at amounts of several ng L⁻¹ or less.^[2] This work demonstrated 50% MA removal by the action of mag-CD microrobots from a 200 ng mL⁻¹ initial MA solution, which differs by three orders of magnitude from the values previously found. In addition, because only 40 min were needed for this achievement, it is safe to say that these microrobots would guarantee a successful performance in real polluted waters.

A “static assay” was then prepared to assess the impact of the magnetic maneuvering on the %MA removal. Thus, the same experimental conditions were kept but without applying a transversal magnetic rotating field. A comparison between the dynamic and the static assays, depicted in Figure 3C, clearly reveals a superior response of the one involving the magnetic motion, with an estimated increment of 56.24%. In this sense, the mag-CD microrobots would act as externally controlled micromixers able to generate an enhanced mass transfer that promotes the MA removal through the establishment of host–guest supramolecular interactions with CD cavities.

We further investigated the impact of CDs on the performance of mag-CD microrobots for MA removal. A “control-blank” sample of magnetic microurchins without CD functionalization was prepared for the experiment and their activity was compared to that of mag-CD microrobots (Figure S8, Supporting Information). Moreover, instead of a transversal magnetic rotating field, an external mechanical shaker was used to contextualize the magnetic navigation of the microrobots. A negligible performance of the unfunctionalized microparticles was observed after 40 min. Thus, the MA capture was undoubtedly related to the hydrophobic cavities of the CD placed on the surface of mag-CD microrobots (Figure 3A). It is important to mention that the percentage yield of MA removal was 99.2 ± 0.1% when mag-CD microrobots were combined with the external shaker. A higher efficiency was estimated in this assay due to the enhanced mass transfer provided by the external shaker. Nonetheless, with real applicability in mind, the intrinsic motion of the mag-CD microrobots can be considered a more interesting approach. As demonstrated in this work, they can produce a mass transfer high enough to decrease MA concentration (ng L⁻¹) from wastewater treatment plants. In addition, lower implementation costs would be needed compared to those associated with the use of external shakers for high volumes of treated water. In terms of the reusability of the mag-CD microrobots, two main factors should be mentioned without considering the stability of the CD-functionalization, which is guaranteed by covalent bonds^{[34, 48]} (Figure S2, Supporting Information). On the one hand, there are the stability of the host–guest complexes between CDs and MA. In this regard, CDs have been reported as a key component of extraction phases for the determination of MA in biological samples like saliva^[34] or urine.^[32] An easy release of the MA from the hydrophobic cavities of CDs has been described in those cases, being that the MA rapidly eluted to be further analyzed. On the other hand, the magnetic motion must be preserved for the mag-CD microrobots to accomplish their final task. SEM images were obtained after the MA removal to address this point. As depicted in Figure S9, Supporting Information, Fe₃O₄ NPs are still observed in the mag-CD microrobots to ensure their magnetic navigation. Consequently, it is clear that the microrobots are reusable.

Finally, the performance of the mag-CD microrobots should be compared with current tertiary or advanced treatments implemented in wastewater treatment plants for MA removal.^[7] This work demonstrates that the performance of the mag-CD microrobots is not limited by specific working conditions, such as pH, as in the case of Fenton reaction treatments. Unlike ozonation, their activity does not involve a transformation to products of higher toxicity and risk for aquatic life, and it occurs in a fuel-free context. Additionally, the magnetic response in mag-CD microrobots also ensures easy retrieval following the completion of their specific tasks, avoiding the separation and recovery issues associated with photocatalytic degradations based on TiO₂. Thus, microrobots provide an eco-friendly approach to MA removal.

3 Conclusion

The emergent environmental threat originating from residues from drugs of abuse is discussed in the context of microrobots and their applicability for water remediation. Mag-CD microrobots, made of a hybrid inorganic hematite/magnetite core together with a surface functionalization with CD, have been presented and used for the removal of MA as the target pollutant. From the comparison of the static and dynamic assays, a synergistic combination of the magnetic motion with the hydrophobic cavities from CDs has been proved, with an estimated enhanced performance for MA removal of 56.24%. Moreover, no adsorption capabilities were detected without CD functionalization. This work highlights how the successful performance of microrobots should be equally attributed to i) autonomous navigation, ii) enhanced mass transfer, and iii) an accurate design involving the functionalization of a specific task.

4 Experimental Section

Conflict of Interest

The authors declare no conflict of interest.