Building a robot with many different components is a challenging task. Even more so, if done at the micro scale. Self-assembly of particles is nothing new. In fact, it has been achieved for decades. Magnetic particles interacting under rotating magnetic fields self-assemble as do components bound together through chemical reactions. Bacterial microswimmers are one such example. However, the end result of these self-assembled micromachines has been very limited – until now.
Scientists researching at the Max Planck Institute for Intelligent Systems in Stuttgart (MPI-IS) took a new approach in self-assembling not one but many different shaped machines only 40 to 50 micro meters in size – about half the diameter of a human hair. They proofed that programmable self-assembly of their micromachines is possible solely through the design and structure of the individual components: by making use of dielectrophoretic forces that evolve around the individual parts when exposed to an electric field. In this environment, the design and structure of the machine frames on the one hand and magnetic actuators on the other allow their controlled configuration, making the assembly programmable.
“We take advantage of the shape- and material-specific forces in a non-uniform electric field,” Alapan explains. “The shape of the machine frame on the one hand and actuators on the other dictates the surrounding gradients. These cause a pulling-force between the units assembling the micromachine. By changing the shape, we control how these gradients are generated and hence how the components are attracted to one another.”
“The components of our micromachines can move relative to each other, which gives another level of complex locomotion,” Yigit says. “Imagine the wheels of a car rotating but the chassis stays unchanged: the car in its entirety moves forward and can go in many directions. Instead of forming rigid connections, each part can move individually.”
Building the individual components was done through a special 3D printing method using two-photon lithography. “Our first design was a microcar, as a homage to the ubiquity of wheeled propulsion in our lives”, Alapan continues. “We fabricated the 3D frame or chassis with its wheel-pockets as this structure generates very attractive gradient forces for the magnetic microactuators – the wheels. Within seconds of applying the electric field, the wheels self-assembled into these pockets!” The researchers then steered the microcar by a vertically rotating magnetic field
Alapan and Yigit tried out many different component sizes and shapes, their tiny self-assembled robots come in many varieties: The researchers were able to build a microcar, a microrotor, something that resembles a small rocket and even a micropump. While it is rotating, magnetic particles at the pumps’ periphery are being moved upwards along its spiral. This causes a pumping effect when one micropump is close to another. The researchers further showed that they can not only assemble motor and structural parts in a configurable way, forming microrobots, but also assemble several microrobots together paving the way for hierarchical multi-robot assemblies.
Being able to move in many different ways is of huge benefit: It could be the deciding factor whether or not such micromachines could one day be deployed in delivering drugs or sense tumor cells inside the body, where a versatile locomotion is key.
The above method of self-assembly to be able to build micromachines of many different shapes and sizes will have a great impact on the scientific community. “Mobile sensing, in vitro targeted drug delivery, single cell manipulation, and precise actuation at this scale are all a great challenge. This new approach could reduce the complexity of these tasks”, says Metin Sitti.
Shape-encoded dynamic assembly of mobile micromachines
Yunus Alapan, Berk Yigit, Onur Beker, Ahmet F. Demirörs & Metin Sitti
Nature Materials (2019)
Director of Physical Intelligence Department at MPI-IS
Tel: +49 711 689-3401
Max Planck Institute for Intelligent Systems (MPI-IS)