Early in my IT career, I spent about a year doing late night tech support for a large insurance company. In spite of the company’s size, it was rare to have more than a dozen people in the building during my shift, so to say that there was a lot of time to kill would be an understatement. One night I walked into the server room to check on something, and the guy who was responsible for keeping tabs on some of the systems was kicked back in a chair watching one of the Terminator movies. I happened to walk in during a scene of a liquid metal terminator robot forming itself into a new shape. The guy who was watching the movie turned to me and told me that what I was seeing on the screen was real. He then tried to convince me that the government already had liquid metal, shape-shifting robots whose capabilities far exceeded those shown in the Terminator movie.
Needless to say, I wasn’t convinced. After all, this was the early 1990s. Microprocessors of the time were still large and bulky, and it definitely would not have been possible to incorporate one of those microprocessors (let alone memory and a battery) into a shape-shifting, liquid metal robot. We actually spent the next hour debating the technological feasibility of a liquid metal robot. By the end of the night, neither of us had wavered from our position. I remained convinced that liquid metal shape-shifting robots were nothing more than a figment of Hollywood’s imagination, and that it would be impossible to actually construct one.
The funny thing is that as many people have noted over the years, Hollywood has a way of inspiring technological innovation. There are countless stories of scientists and engineers building cool tech gear that was inspired by something that they saw in a movie. The movie “Return of the Jedi,” for instance, had a few scenes involving the use of hoverbikes. At the time, those bikes were pure sci-fi. Today, however, the police in Dubai are patrolling the streets on hoverbikes.
Of course, it is one thing to draw inspiration from a movie. It’s quite another to be able to build a real-world version of some futuristic movie prop. If you had asked me six months ago if it would be possible to build a shape-shifting, liquid metal robot I would have said no. Believe it or not, though, creating shape-shifting robots might actually be possible.
So before I delve into a discussion of how such a robot might work, let me take just a moment to delve into some elementary school-level chemistry. A water molecule is made up of two hydrogen atoms and one oxygen atom. In other words, even though water is a liquid, it is actually made up of individual molecules, and each of these molecules has a rigid structure. In other words, even though water molecules are rigid, water behaves like a liquid when it is at room temperature.
The reason why water exists in liquid form at room temperature is because the hydrogen bonds that exist between molecules are really weak and are easily broken.
Now before I go on, consider that under the right circumstances it is possible for solid substances to take on something of a liquid-like behavior if that substance is made up of individual particles rather than being a single mass. Consider the way that the wind sculpts the desert sand into ever-changing dunes, much like wind creates waves in the ocean. Consider how snow can avalanche down the side of a mountain, behaving similarly to a liquid, even though the water that makes up the snow is in solid form.
So with that in mind, consider how an avalanche occurs. Without getting too technical, an avalanche occurs when the top layer of the snowpack dislodges from the snow layer beneath it and then slides down the mountain. Just as broken hydrogen bonds allow water molecules to behave as a liquid, the broken bond between layers of snow allows an avalanche to take on behaviors that are somewhat like a liquid. This concept would most likely be key in building a liquid metal robot. The robot would need to be made up of a huge number of tiny particles that can work together to perform some sort of function, while also having the ability to go between a solid and a liquid state on demand. Needless to say, that’s a tall order; tall, but perhaps not impossible.
Companies like Intel are now manufacturing transistors at near-nanometer scales with each transistor consisting of just a few atoms. Given that it is possible to construct an electronic circuit at such a small scale, it seems plausible that thousands (if not millions) of individual smart particles could be constructed that collectively act as a liquid metal robot.
What’s in that particle?
But what specifically would go into the particles, and what would the particles have to do? At a high level, the particles would have to perform two main functions. They would need to be able to communicate with and coordinate with the other particles (think swarm technology), and they would have to be able to alter states on demand.
As such, the individual modules would require a simple microprocessor, some sort of communications link, and a means of severing or weakening particle level bonds on demand. Bonding could conceivably be controlled with an electromagnet, but the particles would also need some mechanism for articulation. Otherwise, the robot would be unable to move, let alone come together in a solid form.
It also seems likely that some of the particles making up the robot would need to be dedicated to specific purposes. For example, special compute particles might be larger than the other particles and handle most of the robot’s computational requirements.
Of course, one of the biggest challenges in constructing such a robot would be powering it. The use of batteries would be out of the question because each individual particle would need its own battery. Instead, wireless induction power would probably be a better option. Consider for example, how RFID chips do not contain a power source of their own, but rather respond to external signals.
If this approach were to be used, then it also opens the door for a more simplified and elegant means of controlling the robot. Rather than making the robot entirely self-contained, the processing could be offloaded to an external computer that would control the robot by sending signals to individual particles.
Shape-shifting robots: A work in progress
The anatomy and workings of a liquid metal robot as described in this article are pure speculation stemming from my own imagination. Even so, I am not the only one who is thinking about this sort of thing. By pure coincidence, I recently stumbled onto an article describing how researchers at the University of Colorado Boulder are working on shape-shifting robots of their own.
The interesting thing about the university’s approach is that while the robot is not liquid metal per se, it is based on a swarm of tiny particle-like robots. The university based its idea on the behavior of fire ants, which are known to collectively work together to form semi-rigid structures including bridges and ladders.
Featured image: Shutterstock