NASA will start testing a small-scale fusion reactor in about a month in what may be the first step towards building fusion rocket engines that could open the solar system to settlement and tourist traffic.

The Gas Dynamic Mirror (GDM) Fusion Propulsion Experiment is one of several advanced ideas -- ranging from demonstrated to way out -- that NASA is investigating in its constant battle to more efficiently break free of Earth's gravitational grip.

Beyond traditional chemical combustion used by current launch vehicles and the ion drive Deep Space 1 is demonstrating lie several nuclear options. The process of splitting, or fissioning, heavy atoms like uranium to produce energy is being revived. But more energy can be obtained by fusing lightweight atoms into slightly heavier ones -- if you can make it work.

Fusion propulsion was the subject of one of several sessions at the 36th annual Joint Propulsion Conference held this week in Huntsville, Alabama.

"Fusion is a little bit down the road," explained Terry Kammash, a professor of engineering at the University of Michigan who proposed the GDM experiments. "Several schemes are being explored."

Kammash is a veteran of fusion engineering. His experience dates back to Project Sherwood, the initial 1950s effort to harness fusion to produce energy. Along with other scientists, he has seen plasmas -- superhot, dense clouds of hydrogen nuclei -- twist, contort and otherwise defy human efforts to heat and compress them just enough to produce more energy than is pumped into the reaction.

That would seem to make fusion a poor option for a rocket.

"Conditions for building a propulsion system are a little less stringent than for building a power plant," Kammash explained. "If you forget about the money part [i.e., making money], you can build a propulsion plant."

The Gas Dynamic Mirror is an old idea. The idea is to use magnetic fields as mirrors that contain the plasma -- a gas in which the bare nuclei of atoms that have been stripped of their electrons are crammed together by strong magnetic fields. Older mirror devices were plagued by a number of instabilities such as "bad curvature" of the magnetic field that let the nuclei escape.

The GDM, explained William Emrich, the lead engineer on the project at NASA/Marshall, overcomes these instabilities by being long enough to avoid "bad curvature," as well as another effect, "loss cone," that also lets ions wander out.

NASA/Marshall is also investigating a pulsed-propulsion scheme rooted in Project Orion, a system that would have exploded a series small atomic bombs under a rocket with a large bumper plate. Orion was canceled by the ban on nuclear explosions in space. But the concept never entirely went away.

"I think we can actually build a pulse-propulsion system," said Brice Cassenti of United Technologies Research Center in East Hartford, Connecticut. "It's unfortunate that we do know how because the knowledge comes from nuclear weapons."

The promising concept that Cassenti and NASA are investigating, magnetically insulated inertial-confinement fusion, employs a unique combination of fission, fusion and antimatter. The concept derives partly from attempts to have high-energy lasers implode targets of deuterium and tritium (D-T -- the heavy hydrogen and heavy-heavy hydrogen used in most fusion studies) to produce power. Kammash suggested replacing the large, fragile lasers with a quick squirt of antimatter.

Cassenti described the targets that would be used in such a scheme: only 0.8 inches (2 centimeters) across (less than half the width of a ping-pong ball) and just 0.1 ounce (3.5 grams) in mass. Most of the target is deuterium and tritium, with a hollow core and a small chip of uranium 238 to one side. The D-T pellet is coated with uranium to serve as a neutron reflector, and that is coated with tungsten to help contain the blast just for an instant.

In operation, a target is dropped into the combustion chamber and a stream of antiprotons is fired through a pinhole into the core. This triggers fission in the uranium. Neutrons reflect off the uranium shell, and freed electrons form a magnetic field to confine the D-T plasma long enough for a small fusion reaction.

Theoretically, Cassenti said, such an engine could have an Isp of up to 200,000 seconds, although the practical limit is 9,000 seconds -- more than 19 times as efficient as the shuttle's main engines. And unlike most other schemes where high Isp also means low thrust, the pulse propulsion would have a real kick.

Firing at a rate of 136 pellets per second, the pulsed fission-fusion hybrid would accelerate a ship at 1/5 G for extended periods.

"You could get to the inner planets in less than a week," Cassenti said. "This is tourist stuff." Jupiter would take less than a month, allowing "settlement trade."

"We would provide a way to open the solar system," Cassenti said. "The solar system can be settled and you can do trade -- and it can actually pay for itself."

A test of the principle may be just a few years away, Kammash explained. Marshall Space Flight Center is investigating basic methods with a Van de Graff generator that will produce high-speed beams of protons aimed at fusion targets costing about $5,000 each. With this they can learn to position the target and aim the beam.

"Then it becomes really exciting," Kammash said.

Pennsylvania State University is developing a Penning trap, a Star Trek-like magnetic bottle that will hold a small quantity of antiprotons, carefully trapped and cooled after they come flying out of an atom smasher in other experiments. It will hold a paltry 1 trillion or so antiprotons, just enough for one test shot.

"How many times you can do this depends on how many times you can go back and fill it up," Kammash said. But that may be enough to demonstrate the high-octane fuel that space enthusiasts have been seeking to open the spaceways.