Shape Memory Polymers (SMPs)

On 4 October 2004, a crowd stood on the tarmac at Mojave and watched a small white spaceplane called SpaceShipOne climb out of sight for the second time in a week, winning the ten-million-dollar Ansari X Prize for private spaceflight. The craft’s signature trick was not the rocket. It was the tail: on the way down, the whole rear half of the wing and both tail booms folded seventy degrees upward, turning the vehicle into a stable, badminton-shuttlecock shape that needed no piloting through the worst of re-entry. Then, lower and slower, the pilot folded it back flat for an ordinary glide to the runway. The aircraft had, in effect, two shapes, and it switched between them on command.

Plate I

Plate II

That folding tail was pure mechanism: hinges, actuators, a pilot’s hand on a lever. But that same year, a much quieter kind of two-shape trick was being locked down on paper. On 13 April 2004 the United States granted patent 6,720,402 to Andreas Lendlein and Robert Langer, assigned to the GKSS research centre in Germany, for a polymer that could hold more than one shape in its memory and recover each one in turn, entirely on its own, with nothing folding it and nobody flying it.

A Material That Had Already Learned One Trick

Lendlein and Langer were not starting from nothing. Lendlein had shown, while a postdoctoral researcher in Langer’s lab at MIT in the late 1990s, that a biodegradable polymer network could be trained to snap back to a designed shape once warmed. In 2002 the pair published the demonstration that made the idea vivid: a length of suture, tied into a loose overhand knot around a wound, that pulled itself taut into a properly tensioned surgical knot within twenty seconds of being warmed to just above body heat, solving, at a stroke, the awkward business of tying secure knots through a narrow endoscopic opening. Lendlein had by then founded mnemoScience, a spin-off in Aachen, to carry the chemistry out of the lab.

Netpoints and Switching Segments

Every shape-memory polymer network is built from two kinds of chemistry doing two different jobs. A sparse set of permanent junctions (netpoints, made either from true covalent crosslinks or from a stable crystalline or glassy domain that behaves like one) fixes the shape the material was designed to have and never gives it up. Threaded between those netpoints are switching segments: chains that are soft and mobile above a chosen transition and stiff or locked below it. Deform the material while the switching segments are mobile, then cool it back down while still holding that new shape, and the switching segments freeze in place, storing the deformation as a kind of tension the netpoints are straining to release. Warm the material back past the transition and the switching segments go mobile again; with nothing left to hold them, the netpoints simply pull the network back to the one shape it was never willing to forget.

Plate III

What the April 2004 patent actually added to that picture was plurality. Earlier shape-memory networks carried a single switching segment with a single transition, so they could remember exactly one temporary shape at a time. Lendlein and Langer’s claims described a network built with several switching segments, each with its own transition temperature staggered above the last, so that warming the material through each transition in sequence released one stored shape after another. It was a material with, genuinely, more than one thing to remember.

Plate IV

Triggers Beyond a Warm Hand

Heat is the trigger nearly every shape-memory network was designed around, since a glass transition or a melting point gives an obvious, tunable switch. But the same netpoint-and-switching-segment logic has since been rebuilt around other triggers. The following year, Lendlein’s group with Langer showed a network with light-sensitive groups grafted onto the chains: shone with one wavelength of ultraviolet light, the groups link up and lock a stretched shape in place; shone with a different wavelength, the links break and the material springs back. This was a shape-memory cycle run entirely by light, with no warming involved at all. Other groups have built versions triggered by a shift in pH, by an electric field applied across a filler of conducting particles that heats the network from within, and by solvent uptake. Each is the same architecture wearing a different switch.

What the Data Actually Says

Shape-memory behaviour is a design principle, not a chemistry, and it has been built into polyurethanes, epoxies, polyesters and thermoplastic block copolymers alike, which is exactly why this page carries no single density, modulus or transition temperature of its own. Whatever those numbers are belongs to the specific host network a given SMP is built from. The one thing genuinely common to the class is behavioural rather than a value on a data sheet: a sharp, tunable switch between a mobile state that can be deformed and a locked state that holds the deformation, and a reliable return to the one shape the netpoints were built to remember.

Where the Idea Went to Work

Medicine took to shape-memory polymers early, for the same reason Lendlein and Langer chose a suture as their first demonstration: a device that can be threaded into the body in a small, compressed temporary shape and then expand or tighten into its working shape once inside is a device that needs a smaller incision. Cardiovascular stents, punctal plugs for the eye and glaucoma shunts have all been built on the principle. Aerospace engineers picked it up for a different reason entirely: a spacecraft panel or antenna that can be folded flat for launch and then unfold itself once warmed by sunlight needs no separate deployment mechanism to fail. DARPA-funded work on morphing aircraft skins has tested the same idea at a larger scale. Away from either field, the same tunable switch shows up in building-insulation foams, sports equipment and anti-counterfeiting labels that reveal a hidden pattern only once warmed. These are smaller, less dramatic uses of the same trick SpaceShipOne’s hinges performed with a pilot’s hand instead of a molecule.

A Second Shape, On Its Own Terms

The distinction that made 2004 the right year to remember is a fine one, but it is the whole story: SpaceShipOne’s feather needed an actuator and a decision, and the shape-memory network needed neither. Once the netpoints and switching segments are set, the material carries out its own recovery, unassisted, every time it crosses its transition. That is what “memory” means here: not a metaphor borrowed from computing, but a shape a polymer network is chemically incapable of forgetting.