When Crumples Aren't Flaws

Inflatable structures are all around us, from pool floaties and archways at the Tour de France, to life-saving emergency resuscitators and vehicle airbags. As useful as these structures are, they are limited by only having two stable states — inflated and deflated.  

Maybe not for long. A Harvard team comprising engineers and architectural designers have developed a mechanics-based framework for precisely controlling how and where crumples form on inflatable membranes. They used these controlled crumples to create lightweight inflatable objects that change shape after inflation, and that exhibit what physicists call multistability – the ability to lock into not just one, but multiple stable shapes. The research, which was supported by a Department of Defense Multidisciplinary University Research Initiative (MURI) award, lays the groundwork for exploring new kinds of complex, inflatable devices.

The research is published in Advanced Science and led by Katia Bertoldi, the William and Ami Kuan Danoff Family Professor of Applied Mechanics in the John A. Paulson School of Engineering and Applied Sciences (SEAS), and Martin Bechthold, the Kumagai Professor of Architectural Technology in the Harvard Graduate School of Design. The journal featured the research on its inside back cover. 

“This project not only explores the use of multistability in new and perhaps unexpected ways, but it also shows the exciting potential of collaborations between designers and engineers,” Bertoldi said. 

Multistability as both a design and engineering challenge

The project began several years ago when co-first author Hye Jun Youn, then a master’s student in the Harvard Graduate School of Design, started experimenting with thermoplastic polyurethane film – the same material used in most pool toys. She had designed modular inflatable shapes called “PneuBots” that would curiously curl and bend into stable shapes when inflated, piquing her interest in the physical mechanisms behind this behavior.

“I needed scientific data to prove this shape-changing behavior – something I couldn’t prove myself,” said Youn, now a Ph.D. student at the MIT Media Lab.

At Harvard, her search for collaborators led her to structural design professor Bechthold, who became her advisor on the project. She connected with Bertoldi’s group at SEAS through both Bechthold and classmate Leon Kamp, a graduate student in the Bertoldi lab, which was already known for extensive exploration of multistable and shape-morphing structures inspired by both nature and art.

When Youn first brought her hand-held prototypes to a Bertoldi group meeting, then-postdoctoral researcher and co-first author Yi Yang was struck.

“I saw that the inflatable was just a very small pouch or pillow, and I thought, ‘Oh this pouch is actually bistable,’” recalled Yang, now an assistant professor of engineering at Gordon College. “Our lab is very interested to understand how structures change shape, so this triggered my motivation to understand the physics or science behind this bistable structure.” 

Most inflatable membranes are soft and flexible, but when pressurized, they prefer to bend rather than stretch, which naturally produces wrinkles or crumples, like those along the perimeters of Mylar balloons. Crumples are generally considered something to avoid when making inflatables. The Harvard team instead wondered whether those features could be useful. 

Building bistable structures

They started with a flat rectangular pouch and placed notches along its edges. Upon inflation, the notches cause crumples to span the surface, becoming like a mechanical hinge.

“The key idea is whether we can deliberately localize crumples on the surface, so that we can control where these crumples form, and in turn, tune the stability of the whole structure,” Yang said.

By tuning the geometry of the notches and their spacing, the team found they could reliably switch the hinge between two mirror‑image configurations at constant internal pressure, giving each pouch two stable shapes.

Once they understood this basic mechanism, they tessellated many of these bistable units into larger arrays. By arranging the notches along different directions, they could make one‑dimensional chains that fold like a string of hinged panels, or two‑dimensional sheets that snap into a range of three‑dimensional forms.

In smaller lab demonstrations, the team used these programmable hinges to create panels that change curvature or act as simple interactive devices — for example, a small inflatable that toggles a lamp on and off as its crumpled hinge flips.

Scaling up further required a different fabrication approach. Working with staff at the GSD’s fabrication lab, the researchers adapted a large CNC cutting machine to weld intricate patterns into wide rolls of polyurethane. Final edge‑sealing for the largest prototype was completed in partnership with a commercial inflatable manufacturer in Providence, Rhode Island.

The resulting large structure, roughly person‑height and several meters long, can be carried by a few people when deflated, then inflated outdoors into a pavilion-scale shell, which they demonstrated on the lawn of the Science and Engineering Complex. The shell reconfigures between shapes, including a cone‑like enclosure, a tunnel, and a canopy.

“Honestly, I wasn’t convinced we’d be able to demonstrate this at the pavilion scale,” said Bechthold. “Things are not so easy to scale. I think actually demonstrating that there is a spatial component here … I was quite happy with that.”

Potential safety applications

The team also conducted experiments on how multistable inflatable systems could potentially improve safety. Inflatable systems are already widely used in things like airbags and fall-protection equipment. Since multistable structures dissipate energy as they snap between states, the team investigated whether combining inflation with controlled multistability could improve shock absorption.

Drop tests comparing a conventional air cushion with the new multistable design showed that the controlled-crumple inflatable protected fragile objects dropped from higher heights. In one demonstration, a raw egg rebounded and broke when dropped onto a standard airbag but survived when dropped from even higher onto the multistable version, which dissipated energy as its crumple hinge flipped.

The experiments open new design possibilities for adaptive, reconfigurable structures that remain lightweight and easily deployable. Potential applications range from shape‑changing furniture and temporary installations to protective systems and soft robotic devices, the researchers said.

History of collaboration

The project builds on a history of collaboration between the Graduate School of Design and SEAS, in which designers and engineers jointly explore new material behaviors and their architectural implications.

Bechthold, whose earlier work focused on membranes and rigid shells, said the topic felt both personally and environmentally timely. These systems, he noted, “can do so much with very few materials and not a lot of pressure,” aligning with the need for utmost efficiency and design with ever-fewer materials.

For Kamp, who completed a degree in architecture before his Ph.D. in engineering, the collaboration underscores how creative exploration and scientific analysis feed one another. Designers, he said, “just love making stuff, so we just keep making, and if we see something interesting, you make, you make, you make. From the engineering side, you try to analyze and see if we can then unlock new possibilities with structures that you weren’t able to find if you blindly searched that space.”

The research also received U.S. federal support from the Army Research Office grant W911NF-22-1-0219, the Harvard Dean's Competitive Fund for Promising Scholarship, and the Graduate School of Design's Laboratory for Design Technology.