A MICRON-SIZED WALKING ROBOT
Spanning just five microns, a robot developed by Cornell researchers actuates and responds to applied magnetic fields.
Written by Poornima Apte
Diffractive robots that are just 5 microns to 2 microns in size can be directed to complete tasks by controlling the magnetic fields that drive their motions. Image: Cornell University
THERE ARE SMALL ROBOTS and then there are even tinier robots—such as the ones recently developed at Cornell University, which are small enough to interact with (and diffract) light. And, at a mere five microns in size, they are among the smallest walking robots yet.
These microscopic robots have two different magnetic axes with dipoles in opposing directions, allowing for device actuation with a lot more control than a simple one-axis programming system can deliver.
Magnetic axes control also confers the ability to program different conformational shapes into the robot. An out-of-plane magnetic field can get the dipoles to align with the magnetic field, and the panels themselves would contract. Changing the magnetic field can make the conformational shape transform from flat to two-dimensional, pyramidal, and more.
“It gives you control over the shape of the robot, which gives you more applications and more access to doing more interesting things at that scale,” said Conrad Smart, one of the primary researchers on the project at Cornell’s Laboratory of Atomic and Solid State Physics.
Cornell researchers have developed a robot small enough to diffract visible light. These robots can move to local regions—potentially in the body—to take images and measure forces at the microscopic scale. Photo: Jason Koski/Cornell University
OBSTACLES (AND PERKS) OF SCALE
The research team used tools like photolithography, etching, material deposition, and electron beam lithography to define and assemble the robot layer by layer. A chemical process undercuts the substrate the robot is fabricated on, releasing the robot from the surface.
The idea of a microscopic robot is to “disconnect it from the macroscopic world,” Smart said. Doing so enables the robot to measure extremely small things in a controllable and well-defined three-dimensional space. “When you are able to control multiple axes at that scale, you can do a lot of things. You can start to explore drug delivery and other medical applications. You might do some medical imaging with it, you might want to ablate something, and you can do all that innately at the micro scale,” Smart said. “By adding programming, the robot has information embedded in, as opposed to just actuating in one axis.”
Embedded information allows the robot to do assigned tasks like imaging and sensing. The robot can also sense thermal, magnetic, and electrical forces.
“You can place a diffraction grating on the robot and walk it over to a cell that you’re interested in imaging,” Smart said. “You can illuminate the cell through the robot or perform optical functions like focusing and redirecting light. You can extract information about the cell from the optics you have on board the robot.”
A significant advantage comes with being able to bring the optics so close to the sample that “you can build relatively high numerical aperture optics for very low cost. It gives you a way to image very local things with very fine resolution for a very cheap price tag,” Smart added.
“There are challenges to making them smaller, including fabrication, when they get to sizes near the wavelength of light.”
—Conrad Smart, researcher in the Laboratory of Atomic and Solid State Physics at Cornell University
Itai Cohen, professor of physics, in his lab in the Physical Sciences Building. Photo: Jason Koski/Cornell University
MULTIPLE CAPABILITIES
Because of the robot’s very small size, it undergoes Brownian motion—the random motion of an object suspended in a medium—and is flexible enough to sense the thermal background noise of the medium it’s in.
“You can imagine changing this around and building a cantilever on it with a little probe that moves in the fluid, which can be used to measure the stiffness of a cell wall and its thickness,” Smart said.
The robot, especially at the upper end of the size ranges, is also capable of carrying small electronics to detect temperature and perform other kinds of sensing. At times the extremely small size of the robot might make the Brownian motion be too erratic—the robot might simply wander off—but that can be evened out by applying a small countermagnetic field.
Microrobot on the move. Video: Cornell University
The use cases for these robots are vast and they don’t necessarily have to be untethered either, Smart added. For example, they could be used at the end of an endoscope to assist with medical imaging. Larger 50-micron-sized robots are also useful because they need smaller magnetic fields to achieve a fixed motion.
“When you make your robot smaller, you get less real estate to place magnets on your robot and if you get to a size where you can’t put enough magnets, you can’t move your robot around because it won’t respond to the magnetic field you’re applying,” explained Smart, noting that he also envisions multiple optics on the robot, with an array of sensors that will deliver points of information simultaneously.
While the team relied on externally applied magnetics to move the robot because it’s relatively simple, the next frontier to explore would be to make the robots move more innately at the micro scale.
“[Much like an amoeba,] I would love to have a robot that works to locomote in the environment it’s in,” Smart said.
Poornima Apte is a technology writer based in Walpole, Mass.
© 2025 The American Society of Mechanical Engineers. All rights reserved.