Vanderbilt has been perfecting the rocket-powered arm since 2000five years before Darpa kicked off its program to develop better prosthetics for amputees returning from warand has already surmounted a long list of engineering problems. The first was keeping the rocket's exhaust (steam) from scorching the user's skin. The current design does that by internally insulating the reaction chamber, keeping internal temperatures below 425 degrees Fahrenheit and venting the water from its steam exhaust through a porous artificial skin.
"The amount of steam coming through the porous covering on the arm is about the same as a normal sweat on a hot day," said Goldfarb.
The primary safety requirement for the rocket-powered prosthetic was a combustion-free reaction. Goldfarb used hydrogen peroxide as the fuel and aluminum oxide beads coated with iridium as the catalyst. A very controlled aerosol is flowed over the catalyst beads, yielding a reaction whose only by-product is super-heated steam. The steam is used to drive each engine pneumatically at every joint, on demand, with only supply tubes for fuel going from a central tank to each joint. The arm and hand have individual fuel supply lines going to 17 joints.
Each rocket engine in each joint is about the size of a pencil and contains a permanent chamber with the catalyst pellets, which are conserved in the reaction and therefore do not need to be replenished. The hydrogen peroxide fuel produces pure steam to open and close pneumatic valves connected to spring-loaded joints by belts made of the same monofilament used in aircraft. A replaceable fuel tank with a sealed chamber of compressed hydrogen peroxide fits in a compartment in the upper arm to provide one day's energy to power the device (about 18 hours of normal activity).
Whereas conventional prosthetic arms have only a few joints, Vanderbilt's 17 degrees of freedom include a wrist joint that twists and bends, as well as bending fingers and an opposable thumb.
According to Goldfarb, there are no electric motors available that can supply the instant, adjustable power enabled by a rocket-propelled robotic limb. "We use a monopropellant that is activated by contact with a catalyst. Other kinds [require] you to mix two components that ignite, whereas all we have to do is deliver just the right amount of propellant to get just the right amount of instant power to any joint of the arm, hand or finger," said Goldfarb.
Given the regulatory concerns and logistical hurdles compared with traditional approaches, which "use batteries available everywhere," Goldfarb acknowledges the design could be a hard sell for the Dapra program. But if further funding is not forthcoming from that agency, Goldfarb said, Vanderbilt is confident it will be able to continue developing the bionic arm with other sources of government and commercial funding.