Efficient Energy Management for Intelligent Microrobotic Swarms: Design and Impact

Energy serves as the foundational element for all active functions within microrobots. Harvesting devices, such as photovoltaic cells and coils, play a crucial role in converting diverse forms of energy into electricity, while energy storage devices enable uninterrupted operation and liberate microrobots from dependence on external sources. Despite the evident importance of energy, there exists a significant disconnect between the development of energy devices and their integration into microrobotic systems, hampering the deployment of microrobots across diverse fields from agriculture to microsurgery. Here, for transcending the “material for material's sake” focus on simple generic metrics for material optimization, such as energy density, advocating attention to the higher tier transformation of material metrics that determine device‐level performance and so make a meaningful contribution to microrobotic technology is argued. Appropriate metrics for microrobotic swarms challenge the stereotype that tiny energy supplies are impractical; instead, swarms simplify individual microrobotic functions, reducing single energy supply requirements and utilizing swarm energy distribution and management. In addition to essential evaluations of the environmental and social impacts of intelligent microrobotic swarms for their safe and beneficial implementation, research targeting these higher‐tier metrics is crucial to connecting energy material research and microrobot developments effectively.


Introduction
[3][4][5][6][7] Onboard energy is a key enabler for such intelligence even if external energy sources are available.Energy harvesters can contribute to extending the operation time but are restricted by temporal and spatial limitations, which become increasingly problematic as devices get smaller.Batteries are necessary to allow for uninterrupted operation.This applies, for example, to harvesting from electromagnetic radiation, where efficiency drops significantly as devices become much smaller than radiation wavelengths and where, at least for deployment in aqueous media such as inside the human body, penetration depths are minimal for high frequencies (short wavelengths). [8]igure 1.Three levels of development toward powering intelligent microrobotic swarms.Level 0 focuses on the development of materials for miniaturized energy storage and conversion devices.Level 1 emphasizes translating material performance into successful device development.Level 2 explores new opportunities created by swarm activities that broaden the applications of miniaturized energy devices.
However, short-term actuator (or microcontroller) power demands can exceed the deliverable power (energy per unit time) of microbatteries. [9,10]Integration with intelligent energy management and boost circuits will be required to power microrobots. [11,12]This is just one example emphasizing the general need to investigate energy solutions for microrobots within the context of their integration, providing new opportunities and considerations for energy material research.
In fact, most current research in the field of energy devices for microrobots is characterized by the exploration of energy solutions themselves.We define this development phase as Level 0 (Figure 1) because the work largely focuses on finding suitable materials for onboard energy solutions and benchmarking their performance prior to or in preparation for integration.This phase involves understanding the energy requirements and limitations, as well as identifying potential materials and mechanisms to store and convert energy.Such solutions might include harvesting ambient energy, utilizing advanced battery technologies, or drawing inspiration from biological systems.The impact of the research in Level 0 for microrobotics will only follow from research moving one step further, to Level 1, to evaluate their success in powering microscale robots. [13]Thus the focus of Level 1 is shifted toward translating energy solutions into real-world microrobotic implementations.This stage requires developing prototypes and conducting experiments to demonstrate the viability of the proposed energy sources within the context of microrobotic tasks.The performance, efficiency, reliability, and adaptability of the energy solutions in powering these miniature robots become the key benchmarks at this level. [10,14]Beyond this, and increasingly so as microrobots get smaller, lies the enormous potential application area of swarm robotics, which radically transforms energy considerations.Level 2 delves into the intricate dynamics of energy management at the swarm level, marking a crucial transition in the complexity of energy management, where the individual robotic entities in a swarm must collaborate seamlessly through efficient energy management strategies.The challenge lies in coordinating the energy distribution across the swarm to optimize performance and ensure sustained operation. [15]This could involve dynamic task allocation, adaptive communication, and cooperative charging mechanisms.
This perspective underscores the necessity of surpassing Level 0 in energy solution development for microrobotic swarms.While foundational explorations in energy materials at Level 0 are essential, remaining confined within this level, contributing solely to the "materials for material's sake," will not facilitate the deployment either of microrobots or microrobotic swarms, ultimately hindering the progress of the field.It is imperative for the energy research community to widen its focus to include Levels 1 and 2. Level 1 validates the feasibility of energy solutions in real-world single/few robot scenarios, while Level 2 tackles the intricate demands of energy dynamics within the swarm context.By embracing the demands of these higher levels of development, researchers can move the field forward, paving the way for the realization of intelligent microrobotic swarms with efficient energy management.

Materials for Miniaturized Energy Devices (Level 0)
Most current studies investigating materials suitable for tiny energy devices, explore a range of materials to create on-board energy systems and benchmark material performance (Level 0).Though these findings offer a reservoir of possibilities, [16][17][18][19] their seamless translation into device performance remains elusive, primarily due to the isolation of this material development from the intricate integration processes essential for microrobotics.
Take miniaturized batteries as an example.Despite ongoing investigations into current collectors, [20][21][22] electrode materials, [23][24][25] electrolytes, [26][27][28][29] and separators [30,31] within the material-centric paradigm, as depicted in the outer ring of Figure 2, progress in advancing tiny on-board batteries for microrobots remains modest.The principal hurdle lies in the critical technological gap associated with integration.Integrating tiny batteries into systems with limited space proves to be a significant challenge more than metrics of high energy density of materials.Regardless of material choices, enhancing energy storage requires loading more electrode material. [32,33]To achieve this in 3D, beyond in plane interdigitated electrodes, three main strategies are available with static electrolytes: creating a thick electrode layer, overlaying thin films multiple times (winding Swiss rolls or folding multiple layers), and sculpting pillars with a substantial aspect ratio, as shown in the inner ring in Figure 2. [34,35] Strategies with mobile electrolytes involve redox-active material systems that can be loaded in liquid form providing the necessary energy capacity and operating as a conventional redox flow system. [36,37]n additional interface between battery structure (inner ring in Figure 2) and material exploration (outer ring in Figure 2) is evident as depicted by the middle ring in Figure 2.This interface introduces and structures the most important engineering constraints or requirements.For example, creating Swiss rolls and pillar arrays necessitates materials to be deformable and conformable. [38,39]While keeping materials very thin or in liquid form might be a solution in principle, ensuring their stability over battery cycling adds an extra layer of complexity to current research on battery materials.Binders and many other functional additives, often taken for granted in battery research, are often unavailable for tiny batteries.In their absence, maintaining mechanical and chemical stability becomes a challenge because the degradation in any of these two factors will cause eventual device failure.It is also crucial to prevent electrodes from cracking during thickening.[42][43][44] Such fabrication requires that the properties of the initially deposited material remain stable during subsequent processes, such as high-temperature deposition and chemical etching.The ion and electron conduction of battery materials in tiny batteries may present challenges in opposing directions.A thin solid-state electrolyte, for instance, might exhibit lower ion conduction than its thicker counterpart. [29]Electron conduction in electrode materials without conductive additives poses a crucial challenge to using thick electrode layers for achieving high energy density.Besides intrinsic material properties, the miniaturization of batteries faces a primary challenge: the patternability of materials.Advances in materials research that do not consider patternability into small features are irrelevant to the development of economical tiny batteries.It is evident that closing the gap between materials research and battery structure development is crucial before we can effectively utilize advanced materials in tiny on-board batteries.
Challenges regarding the integration of advanced materials into miniaturized batteries for microrobots are reflective of broader issues encountered across all energy devices for microrobotic applications.Identifying the technological gaps that arise at the interface between materials and devices, and find-ing effective solutions to bridge these gaps, is crucial.When we shift our focus from materials to actual devices, research needs to provide clear and measurable performance metrics.For instance, material research aiming to enhance battery energy density should transparently report the actual capacity of the assembled battery, providing a clear definition of area, volume, or weight.Table S1 (Supporting Information) offers a list of representative device-level metrics, including nominal voltage, typical capacity, preferred charge/discharge protocols (charge mode, charge voltage/current, cut-off voltages), maximum/minimum discharge current, maximum pulse discharge current, and cycle life at a given current and voltage.The absence of more devicerelated metrics in material research poses a substantial risk, making it impossible to decide in advance on the final performance achievable by these materials, potentially impeding progress in addressing the energy challenges specific to microrobots.

Energy Devices Integrated in Microrobots (Level 1)
The migration of conceptual energy solutions into tangible microrobotic applications introduces a suite of further material requirements (Level 1).Here, prototypes are developed, and experiments are conducted to validate the viability of proposed energy sources within the practical context of microrobotic tasks.A microrobot's energy can be sourced from external or on-board sources or a combination of both (Figure 3a).External sources include energy harvesters of solar, radio-frequency, vibrational or ultrasound power.On board sources include energy storage through batteries or fuel cells.Regardless of whether the power is directly transferred from the energy harvester, stored in energy storage, or both, it must be adequate to drive electronic functions crucial for the microrobot's operation including sensing, actuation, communication, regulation, control, and any required onboard data analysis.These power requirements set the targets for the Level 0 studies.Figure 3b quantifies the unified relationship between energy, power, and footprint.Generally, the amount of stored energy is E = A × E ft , the product of the footprint (A) of an energy device with the footprint energy density (E ft ).However, E ft is a context-dependent metric at the device level, which cannot meaningfully be replaced by footprints unrelated to device manufacture.To be applicable, Level 0 studies need to address devicelevel metrics rather than material-level performance alone.
We can write the operation time (T) for a specific application drawing average power P with energy source size A as T = A × E ft /P.Typically, one first determines the desired device size (A) and the required operating time (T) for a given application. [8]For example, if we aim for a 1 mm 2 microrobot to operate for 24 h, we identify E ft /P as 2400 (h cm −2 ).This value opens up various design possibilities for the microrobot.On the one hand, reducing the microrobot's power consumption through advanced elec-trical engineering allows for more flexibility in energy device design.If average power consumption drops to 1 nW or less, E ft only needs to reach 2.4 μWh cm −2 , a benchmark achievable with stateof-the-art 1 mm 2 batteries. [45]If functionality at this low power level is sufficient, this already enables the design of autonomous microrobots.On the other hand, increased demands for functions, frequent operations, or enhanced intelligence all lead to higher power consumption.Beyond 1 μW, it becomes challenging for a battery to sustain 24 hour operation, necessitating a combination of energy harvesting and storage.Note that power demands are often intermittent in microrobotics, for both actuators and sensors, so that both the ability to deliver short bursts of power as well as sufficient time-averaged power are important.Figure 3b also indicates that a higher E ft /P value offers greater design flexibility.A low E ft /P value, such as 1000 (h cm −2 ), renders a device smaller than 0.5 mm 2 impractical in many contexts, as it would struggle to operate for more than a few hours, independently.Conversely, an E ft /P value of 10000 (h cm −2 ) opens possibilities for microrobots as small as a grain of salt to be deployed actively, broadening their operational scope to approach the scale of biological cells.To sum up, the design requirements for energy density or power density and the final application are intertwined.
For new energy material solutions to be effective at small scales, research must go beyond exploring materials and their integration or packaging as a single energy device.Since comprehensive integration of energy devices with fluidic systems, pumps, sensors, actuators, communication units, and even data processing units is required, the evaluation of energy solutions at Level 1 involves more device-level metrics, where efficiency, reliability, leakage (self-discharge) and the adaptability of performance under load become pivotal benchmarks.Taking batteries as an example, efficiency might be a metric carried over from Level 0 research, but if this metric is derived from a specific current density favorable for a particular material, its real-world efficiency in applications may be overestimated. [46]daptability, often overlooked in Level 0 research, becomes crucial at this stage. [48]Insufficient pulse power can impede the activation of electronic functions, limiting the utility of batteries in microrobots.The energy source must accommodate more than 1000-fold changes in a current draw between the sleep and active modes of microrobots. [10]Moreover, different harvesting sources exhibit varying optimal operating points that change with harvesting conditions.For example, the open circuit voltage of a solar cell may double when transitioning from indoor lighting to sunlight.In contrast, a microbial fuel cell delivers a voltage fluctuating between 500 and 800 mV.These varied requirements call for a power management unit capable of seamlessly transitioning between active and sleep modes, while also engaging the energy harvester to provide additional energy or recharge the battery.In this case, providing a precise report on attainable voltage at different currents becomes even more crucial than simply reporting capacity because the design of the power management unit is based on monitored voltages at different currents.
Critically, inflating capacity reporting artificially, by including the capacity below 0.5 V, is misleading for the useable capability in most applications.Additionally, the temporal high-power or high-voltage demands imposed by actuation pose substantial challenges.For instance, an ionic expansion bending actuator like polypyrrole on gold may involve switching in millisecond timescales and hence peak current demands for microrobots up to the milliampere range.Similarly, a piezoelectric actuator typically requires an electric field of 1 MV m −1 . [49]Considering a 100 μm thick actuator, the voltage requirements would be 100 V. Fulfilling these high voltage demands necessitates the implementation of a boost circuit.
Distributed energy sources within a single microrobot can tackle the challenge of meeting diverse power requirements.For instance, down-converting the operating voltage of a lithiumion battery, which typically exceeds 4, to 0.5 V for electronic operations poses a challenging conversion ratio.On the other hand, aqueous batteries can operate below 1 V, offering a solution for more efficient down-conversion or direct power for electronics.This concept extends to the integration of energy harvesters with varying voltage profiles, ranging from 0.5 V to over 3 V. Figure 3c illustrates a 3D stacked arrangement of multiple batteries using advanced packaging technologies. [47]Each battery can serve different functions or collectively meet a large energy demand.Although this design appears promising, practical demonstrations are yet to be realized, demanding research beyond Level 0 that shifts focus to packaging, integration, and functional aspects in an integrated form.

New Opportunities and Challenges with Microrobotic Swarms (Level 2)
As energy supply equipment is scaled down, the E ft values typically decrease, since the thickness of energy devices must also substantially decrease, influencing the linear lines in Figure 3b and making the extrapolation to 0 impractical.Many physical phenomena change as devices become smaller, such as the relative magnitudes of surface tension (surface energy), gravitational potential energy, kinetic energy (inertial forces), and viscously dissipated energy.Consequently, implementing the functions of a microrobot with macroscopic mechanisms, when transitioning down toward biological cell scales, can become inefficient or impossible.Indeed, it may become necessary to distribute functions and tasks collectively between modules when they become so small.
Consider for example a swarm of microrobots, each equipped with a high-accuracy sensor, capable of precise localization and global mapping.The energy consumption required for each microrobot to achieve precise global mapping of a terrain or even its current location is undoubtedly high, making its usage for this task intractable.Instead, the swarm behavior of many microrobots, coordinating their actions to achieve a common goal, can enable such tasks to be addressed using microrobots each with limited individual capabilities of sensing, actuation, communication, and computing. [15,50]Swarm solutions, Level 2, represent a critical advance at microscopic scales, within which multiple microrobotic entities may form user-defined assemblies to accomplish various tasks.For localization and mapping tasks, the collective solution process can be guided by simply measuring the distance to a neighboring robot rather than relying on heavy communication between all robots and individual high-level intelligence (Figure 4a). [51]Dynamic shapes can be formed by comparisons between neighboring robots (a form of distance geometry, illustrated in inset in Figure 4a). [52]Even without external ordering factors, tiny entities can form a cluster by intrinsic dynamic feedback between speed and density. [53,54]Moreover, cohesive and communicating swarms are able to avoid harmful spots even if only one other individual senses the danger. [55]Such principles are key for the design of energy-efficient robotic swarms, as they determine the minimum number of active scouting robots to safely guide a swarm.In swarm design can thus differentiate, if an orchestrated, intermittent scouting by all individuals or diversification with specially designed scouting units is more energy efficient for a given task.[56] c) Swimming batteries controlled by a magnet and controlled assembly in the parallel form.Data are reproduced with permission from ref. [57]   Note that not all energy-intensive tasks for microrobots must necessarily be powered by onboard electrical energy even for autonomous microrobots.This is particularly important for energy-intensive actuation and especially locomotion tasks.In some cases, it is sufficient for locomotion to be externally controlled, decoupled from the microrobot itself.Figure 4b illustrates that the movement of microscale robot swarms can be decoupled from on-board electronic functions by using external magnetic control. [56]In the many cases where autonomous control of locomotion is preferable, allowing true mutually interactive autonomous individual and swarm motions, external energy sources can still be employed directly without conversions to electrical form.On-board power is then only needed for switching the coupling of external energy fields to actuators.For example, helical microrobots can switch their rotational motion response to external magnetic fields by tuning and detuning their geometry (e.g., stretching and compressing the helix) to match or unmatch various driving frequencies.Similarly, ultrasound-driven microrobots can be externally locomoted by choosing resonant frequencies of ultrasound drivers. [58]In both these cases, autonomous behavior can be reached by powering the local electronics and tuning actuators, along with sensors for feedback control, without needing to provide much larger powers for locomotion.Similar considerations apply to exerting large bending forces or torques within microrobots, which may be required for example to penetrate tissue or perform surgical operations.Instead of an extremely energy-consuming way to stretch or compress bonds between atoms, line defects such as dislocations, that move and make the material plastic.A similar concept might also work for microrobotic swarms for morphing a collective shape.
Such swarm activities open new possibilities for energy source design and management.
Since the function of an individual microrobot entity can be reframed, reducing the value of P, the footprint limit of E ft can be bypassed in such cases, maintaining the operation driven by on-board energy storage effectively.Efficient and precise energy utilization is crucial for the operations of each robot.Integrating an energy source without considering their actual energy needs would lead to wasteful behavior in the swarm. [59,60]In addition to the above examples, when resources, such as chemical fuels for fuel cells, sunlight for solar cells, and redox species for redoxflow batteries are known to be near the task, a robot with large energy storage might be unnecessary.Harvesting is even possible from the drag forces at the boundary with the surrounding fluid may be employed for contactless recharging in the future.In such cases, the robot could utilize available fuels to complete tasks.Conversely, robots facing challenges in resource-less areas would prioritize relying on their on-board energy storage.But even in these cases, swarm activity may also involve intelligent energy sharing.
Multiple energy sources can be interconnected to meet the variability of energy demand.For instance, magnetically controlled swimming batteries can be connected in series and parallel forms, [57] as depicted in Figure 4c.The distinct emphasis on electric specifications between electronics and batteries is evident here.Battery engineers often focus on battery capacity, particularly when reporting the increased capacity of the parallel connection.However, this result does not necessarily indicate a multifold improvement in attainable current for such configurations.Without clear data on the current of parallel-connected batteries, it is challenging to design a swarm capable of performing complex tasks requiring more energy than a single energy source can provide.More generally, the development of movable tiny batteries that are capable of assembly into an energy-dense power bank align with the development of swarm microrobots because both fields need to address the issue of connection or communication between neighboring entities.
Based on dynamic docking connections between different entities, the energy management in a swarm can be decentralized and organized in parallel with potentially differentiated microrobots.Foraging robots with energy harvesting capabilities will seek out relevant resources, while other functional robots will perform user-defined tasks.The energy awareness of individual robots is the key to the efficient use of overall energy in the swarm.Such strategies are heavily reliant on Level 1 development, where energy awareness needs to be integrated with on-board energy sources to obtain precise information on the residual energy levels.Energy management strategies for microrobotic swarms begin at the design stage, where tasks and working environments are defined to generate suitable algorithms for distributing energy among the swarm.An advantage of swarm technology is its ability to incorporate redundant energy sources as contingency plans for energy source failures.Furthermore, swarm design can incorporate specific charging docks to unlimited sources to replenish energy-limited devices, thereby mitigating the risk of energy source failures.
Consider, for example, an early warning system for a complex condition like disease, for which false positives can occur using simple low-energy budget tests, requiring much larger resources to report accurately.Microrobotic swarms can initially forage for positive early warning signals autonomously, using external energy for locomotion and on-board energy for navigation and control.On positive detection, they can signal to attract nearby microrobots, first to verify the warning, second to convey local messages to specialist sensor microrobots in the vicinity for indepen-dent test verification, third to assemble energy from other microrobots or resource modules to support longer range communication or locomotion.The latter can involve assembling larger communication structures like antennae to allow warning signals to be conveyed to remote information systems (e.g., outside the body in the case of human disease, or at the farmhouse in the case of crop disease, or to the surface in the case of earthquake disaster search and rescue).Given the limited energy resources at small scales, the first development focus should be placed on energyefficient hardware design, which optimizes the hardware components (such as sensors, actuators, and communication modules) to minimize energy consumption and uses low-power microcontrollers.Additionally, advancements in energy harvesters to harness ambient energy sources and the integration of batteries with enhanced energy density are essential.Dynamic task management systems should be developed, accounting for prioritized tasks and energy-intensive operations during periods of high energy availability.Energy-aware algorithms that balance task accuracy and energy utilization are essential to distribute tasks in a predictive manner.Furthermore, the swarm necessitates docking strategies to facilitate recharging.These docking stations serve as vital nodes where microrobots can replenish their energy reserves, ensuring sustained operation and minimizing downtime.
In Level 2, the generic challenge lies in efficiently distributing functions across the swarm to optimize performance, energy use, and ensure sustained operation.Energy awareness and communication between different entities emphasize the importance of scalability and consistency of integrable energy sources, which is often overlooked in Level 1, where research has focused on the functional implementation of individual devices.Level 2 work serves as a direct response to a crucial question that remains unaddressed by Level 0 or Level 1 research: "When are miniaturized and integrable energy devices necessary for swarm microrobots?"Providing answers to this question is pivotal for establishing and advancing the field of miniaturizing energy harvesting and storage devices.

Perspectives
Millibots and microrobots typically span from centimeter to micrometer scales.Within the centimeter to millimeter range, allin-one microrobots with integrated energy supplies and multiple functionalities have already proved feasible.Coin cells have been used in many millibots and thin film batteries have found applications in tiny computers through stack integration (Figure 5).In this scenario, a singular microrobot can execute complex tasks, necessitating energy supplies capable of meeting diverse demands from functional components, including short-term highpower requirements and uninterrupted operation.The focus of materials development then shifts toward device-level metrics and successful demonstration.As microrobots continue to shrink to biological sizes for tasks such as precise drug delivery within the body, monolithic integration may prove to be impractical for creating fully functional robots.Moreover, manufacturing tools for large-format batteries are not available.Instead, microscale technologies such as photolithography, 3D printing, and self-assembly become essential for developing microbatteries with the target of more than one milliampere hour per square Figure 5.A generalized design paradigm for energy supplies for microrobots.From centimeter to millimeter scales (the transition point depends on the technology and may decrease to below 1 mm), monolithic integration has been achieved to form all-in-one devices capable of performing complex tasks, while further miniaturization can rely on distributed energy sources and functionality to form swarms that perform complex tasks.
centimeter (footprint of the full battery).Battery chemistries need to be carefully selected to minimize the energy loss over the up or down conversion.For instance, aqueous batteries are suitable when the voltage requirement is below 1.5 V, while electrolytes stable within a wide potential window are preferred for higher voltage needs, as in lithium-ion batteries.A paradigm shift toward intelligent swarm activity emerges as one means to overcome inherent barriers for both energy supplies and smart functionalities at the micrometer scale (Figure 5).In this context, simplified robots with limited functions and energy supplies can be designed, while the collective behavior of numerous such robots, whether assembled as in [61] or in swarms, efficiently addresses energy challenges and enables the execution of complex activities.Larger numbers of robots do increase concerns about environmental impact, even if their total mass is small, necessitating energy devices that pose minimal risk of harming the environment.
The deployment of microrobotic swarms in various applications necessitates a critical evaluation of potential domains to discern their tangible benefits.Applications such as precision agriculture and infrastructure inspection demand the capabilities of microrobotic swarms to navigate through unpredictable and expansive environments. [62,63]Static sensor networks require orders of magnitude more reporters for the common case of moderately slowly changing and sparse events (like disease) than mobile robotic sensors.Decentralization proves advantageous for efficiency through parallel and coordinated operations, particularly in the context of disease outbreak identification and infrastructure defect detection.Future research should prioritize strategies for complex feature comprehension through information fusion among diverse robots, [64] paving the way for decentralized collaborative activities.The appeal of microrobotic swarms extends to fault-tolerant systems resistant to malfunctions.Yet, the human component remains pivotal, necessitating advanced human-system interaction strategies for effective deployment.Crucial considerations for safety and security are emphasized in domains where microrobots must confront emergencies without external infrastructure.The challenges posed by limited energy and computational power find a solution in microrobotic swarms with redundant systems, catering to maintained performance in the event of a failure.The prospect of precision medicine presents a major challenge, most efficiently addressed with active mobile intelligent agents, with on-board sources of power, which allow coordination among numerous robots with limited capabilities.
Despite unique benefits, the potentially large quantity of these tiny robots in the environment raises additional concerns.Indeed, environmental impact is a critical aspect that demands attention in the advancement of microrobotic swarms.While these miniature robots hold tremendous potential for revolutionizing various industries, the unintended impact on the environment needs to be precisely investigated and mitigated before being applied broadly.The production of electronic waste, energy consumption, and potential toxicity of materials used in the fabrication of microrobots as well as their recovery and recycling to avoid waste and long-term accumulation in the environment all need to be addressed in the design process, and this is especially true for materials and processes used in energy powering of devices.
[67][68] First and foremost, the biodegradability or recoverability of materials used in microrobots needs a thorough assessment.Additionally, the environmental toxicity of these materials in terms of agglomeration and solubility behavior as well as abiotic changes should be evaluated to understand their potential impact on ecosystems.Unfortunately, existing guidelines and protocols for testing ecotoxicity, for instance of the micro-and nano-plastics which might be used in microrobotic systems, are notably absent.Standardization concerning chemical transformation, loss of surface coatings, bonding of other substances etc. are lacking but strongly needed in extended exposure models. [69,70]Moreover, the interaction of microrobotic swarms with the environment remains poorly understood.Life-cycle assessments, while providing insights into specific scopes, often lack long-term perspectives. [71]nderstanding how microrobotic swarms might influence and interact with the environment over extended periods is crucial for making informed decisions about their deployment and developing the most sustainable version of microrobotic swarms from the early beginning.
Social concerns are also important in the integration of microrobotic swarms into society.Privacy, safety, and security issues arise as these swarms are equipped with sensors capable of collecting data. [72]There is the potential for misuse, including unauthorized surveillance, espionage, hacker attacks, and even sabotage.Regulations governing the operation of microrobotic swarms in public spaces are essential, requiring not only technology assessments but also clear risk level identification and legal frameworks.Ethical considerations are paramount in the development and deployment of microrobotic swarms.Questions about the ethical treatment of these robots, their potential for autonomous decision-making, and how they interact with humans need careful examination.Ensuring that microrobotic swarms adhere to ethical standards and avoid unintended consequences in their deployment is crucial.Technological characteristics that secure trust and responsibility in microrobotic swarms will rely on precise technology assessments.Establishing different risk classes in terms of legal, ethical, and environmental impact is essential.This nuanced approach recognizes that not all microrobotic systems carry the same level of risk and impact, allowing for targeted regulatory measures.
In conclusion, while the research advances in microrobotic swarms promise groundbreaking applications, a comprehensive approach is needed.Such a responsible approach can only seriously commence for this huge potential market for energy materials, when one has identified, as this Perspective proposes, the three levels of research integration required in developing suitable material energy solutions.As our final discussion has argued, recognizing the undeniable relationship between applications, research challenges, and environmental and social impacts is crucial.A collaborative effort between researchers across diverse domains is envisioned to propel innovative developments and shape the future agenda of microrobotic swarm research.Balancing technological progress with ethical considerations, environmental sustainability, and societal well-being is the key to unlocking the full potential of microrobotic swarms responsibly and beneficially.

Figure 2 .
Figure 2. Battery development transition from materials to devices, and the critical interface between them, where challenges must be identified and addressed.The outer ring depicts essential components necessary for assembling a battery.The middle ring outlines the challenges associated with integrating materials into small-scale devices.The inner ring illustrates three battery architectures capable of achieving high energy density.

Figure 3 .
Figure 3. Energy solutions and requirements.a) Two different energy solutions for microrobotic functions.b) The relationship between energy density, power, footprint, and operation time as a guide for the design of energy devices.c) Distributed battery design on a chip by advanced 3D packaging technologies.Data are reproduced with permission from ref.[47]

Figure 4 .
Figure 4. Robotic swarm.a) Self-assembly of hundreds of centimeter-scale robots and collective formation of a user-defined shape.Data are reproduced with permission from ref.[51] b) Microscale robots forming into a swarm.Data are reproduced with permission from ref.[56] c) Swimming batteries controlled by a magnet and controlled assembly in the parallel form.Data are reproduced with permission from ref.[57]