Skylar Tibbits has some radical ideas about the future of manufacturing.
"If you look at how things are manufactured at every other scale other than the human scale—look at DNA and cells and proteins, then look at the planetary scale—everything is built through self assembly," he says. "But at the human scale, it's the opposite. Everything is built top down. We take components and we force them together."
Tibbits is a research scientist in MIT's department of architecture, but his title belies the radical nature of his work. At MIT, he's developing materials and objects that can be programmed to assemble themselves. In 2011, he set up a lab to experiment with 4D printing, a process that uses 3D printers to produce material that will grow and change on its own. Since then, the Self-Assembly Lab, which Tibbits runs with co-director Jared Laucks, has been experimenting with a series of materials that can be "programmed" to self-construct. They've come up with flat-pack furniture that can build itself, for instance, as well as the textile that could make self-lacing sneakers possible.
Now, in collaboration with designer Marcelo Coelho, who runs an eponymous design studio in Cambridge, Tibbits and his team are applying their work to consumer electronics.
Still in its infancy, their new project explores how with a few components, a source of energy, and the right interactions, a cell phone could "build itself," without the need for human or even high-tech automation. Not only does the project demonstrate Tibbit's research into self-assembly on a practical level, it could also have major implications on manufacturing.
The original impetus for the project was MIT professor David Mellis's DIY cell phone, which began as a series of open-source instructions for a basic cell-phone design using only $200 in parts. Mellis made the project into a kit, which got Tibbits and his team thinking: "What are the components of this, and how can we make it assemble itself?"
Fast-forward a couple years from when Mellis's phone was released, and Tibbits and his lab have a rough prototype for the self-assembled cell phone. It's composed of six parts that assemble into two different phones. The parts are put into a tumbler (think of something like a cement mixer) and tossed around until the parts click together into the phones. Depending on the speed of the tumbler, it can take under a minute. Here's a video showing it in action:
It looks exceedingly simple—and it is much simpler that most automated assembly techniques—but there are a number of complex ingredients involved that make it possible. First, the speed that the tumbler rotates has to be fast enough so that the components are tossed around, but not so fast that they break. The components also have a series of lock-and-key mechanisms that allows the proper connection between parts, and blocks the wrong ones. Lastly, there needs to be something that makes the parts "stick" once they connect, like adhesive or Velcro. In this case, Tibbits and team used magnets of varying polarity, so that only the parts that are meant to connect are strongly attracted to each other.
Tibbits and his team have been working on the concept since 2013, and have been experimenting with it and refining the process for roughly a year. Right now the process is low-budget and simplified, but what's most exciting to Tibbits is that it's completely feasible to imagine scaling it up for mass production. All of the ingredients detailed above can be optimized: You can play with different tumbler speeds, for example, or throw a ton of parts in the mix for faster assembly. At the lab, they added a foam lining to the tumbler so that the parts don't get banged up in the assembly process, but Tibbits says there's room for improvement there as well. As most factories already use tumblers for sifting through materials and separating out debris, their wouldn't be a lot of added cost for equipment.
In terms of manufacturing, Tibbits says that companies have been trying to keep costs low in one of two ways: shifting labor overseas or automating it. Either way, it's eliminating jobs. The self-assembly method wouldn't do much to save human labor, but it would automate the process at a much lower cost.
The most exciting use for their process might be in terms of design. "Right now the phone is predetermined, and we’re using this process to assemble that phone," says Tibbits. "But imagine you take a circuit board and you have different logical building blocks and those logical building blocks can be tumbled around—you can have different functionalities." For example, you could take several of Google's project Ara phones, put them in a tumbler, and allow different kinds of phones to emerge. "Essentially the holy grail is [that] you want complete design freedom," Tibbits says. "So whatever you want to design or make, you have complete freedom in that space with the minimum number of unique components. We’re just scratching the surface."
Correction: An earlier version of this story noted that the Self-Assembly Lab has received funding from DARPA. Skylar Tibbits received funding from DARPA as a grad student, but not for the lab. The article has been updated to reflect that.