快猫短视频

In the next industrial revolution, will we manufacture goods, or will they construct themselves? Michael Reilly sits down with architect and computer scientist Skylar Tibbits to talk about 4D printing, programmable matter and the future of things

You run the Self-Assembly lab at the Massachusetts Institute of Technology. What does 鈥渟elf-assembly鈥� mean?
In general, self-assembly is a process by which disordered parts build into an ordered structure through simple interactions with each other. For example, we worked with Arthur Olson at the Scripps Research Institute to build a model of the polio virus. If you put the pieces of the model in a glass container, they just sit there. But when you impart some energy 鈥� shake the glass 鈥� they come together. Shake harder, and it comes apart again. You have these different states based on how much energy you put into the system; you don鈥檛 control where every piece goes, it assembles itself.

You have combined smart materials with 3D printing to create 鈥�4D printing鈥�. What is that?
At first, the name was a bit of a joke. When you 3D print an object, it鈥檚 a fixed, static thing. If you want something more complex, you need to print it as parts and then assemble it. We thought, instead of assembling intelligence into it afterwards, why not print intelligence into it?

鈥淭he problem is that when you 3D print an object, it鈥檚 a fixed, static thing鈥�

We collaborated with 3D printing firm Stratasys, which developed a printing material that expands 150 per cent when it meets water. That let us print objects that can change over time. Printing isn鈥檛 the end; it鈥檚 just the start. That鈥檚 why we call it 4D printing. We have done a few prototypes so far 鈥� printed a tube that, when you submerge it in water, folds itself into a cube. People are really excited about it, but I think we have barely scratched the surface.

A lot of people see 3D printing as the start of another industrial revolution. Do you agree?
Across industry, academia and government everyone鈥檚 looking at it, but there are still a number of road blocks, including the speed of printing and limitations of materials. Current materials can鈥檛 compete structurally with our existing processes. We need smarter systems and custom algorithms 鈥� that鈥檚 what the 4D printing was specifically targeting.

You have said size limitation is also a big hurdle.
We can鈥檛 print big things yet. The example I often give is that to 3D print a skyscraper, you would first have to build a skyscraper-sized printer. We are trying to address this with Hyperform, a project in which we use mathematical programs to fold objects so that you can print them on a small printer. Then you can assemble them 鈥� by unfolding and snapping pieces into place intuitively 鈥� so they are much larger than the printer that made them. But that鈥檚 still at an early stage.

I think 3D printing has huge implications for many industries, but we have to overcome these hurdles. I have no doubt that we will, but right now it鈥檚 a bit hyped.

There is a lot of research into self-assembly on the microscopic scale. But you are working on things we can see, handle and even live in.
Since the industrial revolution, we have become really good at making things at the human scale, from cars to clothes to consumer electronics. We won鈥檛 try to beat those manufacturing processes. Instead, we will look to outlying scenarios: disaster relief, and areas where we can鈥檛 build today because it鈥檚 too expensive, dangerous or hard to get to. We are looking at space construction, infrastructure repair and control systems, smarter consumer products and US manufacturing opportunities that historically have been outsourced.

Will you work on more traditional buildings?
We are doing research and design for larger objects. Building construction is much less efficient and more energy-intensive than most manufacturing, partly because it鈥檚 a one-off: you only build a building once. It鈥檚 also dangerous. So it鈥檚 an area where there鈥檚 room and interest to develop a new system.

What might that new system look like?
We haven鈥檛 developed applications for construction yet, so it鈥檚 hard to say. But I think it would focus on specific systems on the construction site, rather than everything from the ground-up. That could be materials that adapt to climate, user demands or loading conditions; connectors for programmable earthquake response; or components that inform the contractor where they should be placed and how they should be connected. They might even analyse physical constraints that the user cannot see or feel, and that would be hard to compute externally.

How could smart materials improve large-scale systems, like water infrastructure?
You might have a smart pipe that carries water around a city that, in one state, could close off to become a valve. So it pinches to create higher pressure or to shut off the stream. Or it could expand to allow more capacity, or undulate to move the water, like peristalsis 鈥� the wave of muscle contractions that moves food through our digestive tract.

Do you look to the natural world for inspiration when designing smart systems?
Obviously we are inspired by it, but I am sensitive about not taking the argument that biomimicry is inherently good. I don鈥檛 want to be so literal about the idea of mimicking nature, because that鈥檚 a trap: if it looks like nature, the thinking goes, it鈥檚 good. Or if I just scale it to a different environment, then it will work.

That鈥檚 not necessarily true 鈥� many of the functionalities that nature has built in have evolved for specific reasons under super-specific conditions. Take the grippers of geckos鈥� feet. That may be a good solution for sticking to things and climbing, but not for certain surfaces or applications. Nature is not always optimal, and it may be a terrible solution for certain environments.

Are there other areas where you think smart materials could make a difference?
Smart materials could make improvements across aerospace, the marine industry and the automotive industry as all of these rely on shape and mechanical actuation for performance.

As you drive a car, for example, your performance, the road conditions and the environment are all changing. If you have materials that can respond to that 鈥� better tyre grip as you go through turns, change in tyre shape to channel water away 鈥� that means performance and safety can increase. Or for an athlete, your shoes or clothing could change in response to your body temperature or sweat.

Smart materials could be used for anything where shape or material properties could change, but instead of forcing them to change mechanically, you do it through the materials themselves.

Do you think smart materials will change the way we engineer systems more generally?
The normal engineering approach has been to throw a ton of money at problems to create 鈥渋ntelligent鈥� super-engineered systems. My hope is that over time things become more elegant, so materials fundamentally respond as they need to.

What does all of this mean for the future of making things?
The production of components is already fairly well automated, as with industrial robot arms that weld things together, but there are other things that are much more difficult to automate. If we really rethink how things come together, it could make manufacturing more efficient. Automation could mean more distributed systems dump things in one place, then we add energy 鈥� agitate, shake, whatever 鈥� and they come together.

This article appeared in print under the headline 鈥淭he future will build itself鈥�