“Building robots that operate independently at sizes below one millimeter is incredibly difficult,” it is exceedingly hard to create robots that function on their own and which are less than one millimeter in size, a problem that has defied engineers over the last 40 years. This obstacle has been already overcome with the production of fully autonomous, programmable robots only 200 × 300 × 50 micrometers, or the size of a grain of salt, but able to sense, compute and move in fluid environments over months at a time.
The size of biological microorganisms presents special physical limits to operation. On human scales, the forces of gravity and inertia are overpowered by the forces of drag and viscosity; it becomes easy to move through water like trying to push it through tar. The common locomotion strategies: legs, arms, or fins, are unpractical, fabrication-wise as well as in terms of permanence. The breakthrough propulsion system that is invented in Penn engineering avoids these limitations altogether. The robots cause the movement of ions in the liquid around them by creating an electric field between the integrated electrodes. The transfer of momentum to neighboring water molecules by these ions forms a controllable electrohydrodynamic flow propelling the robot. It is similar to EHD thrust systems investigated in centimeter-scale aerial robots, but does not contain moving components, which makes it robust and easy to fabricate.
Energy budgets are minute at this microscale. The photovoltaic cells of the robots only generate 75 nanowatts, which is more than 100,000 times less than a smartwatch. The degraded performance of such constraints necessitated application-specific low-leakage CMOS circuits and a small instruction set architecture that reduces complex propulsion or sensing programs to a single command. A prototype of the onboard electronics created by the team at the University of Michigan, with the largest computer in the world, entails the integration of processor, memory and temperature sensors in less than a millimeter of space. The robots are powered by pulses of light through which programming is achieved, and each unit has a special optical address so that it can be given a specific task.
Such a close integration allows complex behavior. An example is a robot with a resolution of 0.3 o C to measure the local temperature, and signal it by waggling its propulsion in a bees-like communication protocol. This sensing ability is directly applicable in biomedical applications, in which the temperature gradient can be used to indicate cellular activity or inflammation. Similar microrobots, with local biochemical or thermal responses, might be used in targeted drug delivery, analogous to magnetically controlled nanorobots which have successfully delivered fluorouracil to tumors in vivo.
The principles of propulsion physics mirror in soft robotics EHD actuators, where the forces of ion-drag, dielectrophoretic forces and electrostriction interact to produce complex flow configurations. Under larger EHD devices, the electrode geometry and voltage control can be carefully selected to provide thrust densities of more than 13 -1 m 2, and efficacy of about 3 -1 mN/W. The Penn robots take advantage of similar ion-fluid coupling at the sub-millimeter scale, at orders of magnitude reduced power levels, and circuit simplicity and electrode durability is of paramount importance. They are lithographically patterned using platinum as their electrode layers and are fabricated with the CMOS die and the entire assembly is etched out into solution, with yields of over 50%.
The other frontier is the coordination of more than one unit. The pulse-coupled CMOS oscillator has also been studied, showing that microscopic machine actuation phases can be synchronized, generating metachronal waves that resemble ciliary motions. This kind of coordination would enable swarms of these salt-sized robots to combine fluids, carry microscale-scale payloads or even put buildings together within the body as a whole. The address and optical programming channel of each robot allows differentiated roles in a swarm already; with complex multi-agent tasks, no physical tether or centralized control, this technique can enable swarm robotics to perform.
At the present, the robots are working in hydrogen peroxide solution that cannot be directly used in the medical field. The team is in progress of designing biocompatible actuators which are comparable to the electrical properties of the current design and this will lead to applications in physiological systems. The adapted versions would be able to travel through the blood plasma or even interstitial fluid, using their high accuracy sensing capabilities to detect tissue health, deliver drugs, or aid in microsurgeries.
The approach is scalable in manufacturing: Lithographic processing enables massive parallel manufacturing, and with scale, the price per robot is a cent. Control hardware is also open-source; the group at Miskin has shown an example of a $100 system with normal LEDs, a Raspberry Pi, and a smartphone macro lens which is comparable to microscopes in the laboratory. This microrobotics democratization of the high-functioning could hasten results in other areas such as biomedical engineering and the environmental sensors.
The integration of ultra-low-power computing, robust electrohydrodynamic propulsion, and scalable micro fabrication using these sub-millimeter robots is a developmental breakthrough. They are in a position to become proactive in spaces that engineered systems have, so far, found impractical to explore, be it sensing the environment in cellular realms to detect disease or arranging assembly lines in fluidic chambers on a micro-scale.
