Applying for NASA Innovative Advanced Concepts (NIAC) 2027 Grant with Project StarSplice Direct Fusion Drive
A while ago, the Princeton Field-Reversed Configuration reactor was invented by Samuel A. Cohen, which uses a linear magnetic confinement system to generate thrust by nuclear fusion - this was the inception of the Direct Fusion Drive concept. Since then, companies like Pulsar Fusion through their Sunbird program have attempted to commercialize this technology for deep space flights:
After reviewing this I decided to try my hand at devising a Direct Fusion Drive with an alternative architecture, and I decided to reach out to Dr. Egedal at the Department of Physics at UW-Washington.
Dreamscape Systems is studying a new kind of propulsion concept called Project StarSplice to be submitted to NASA's 2027 NAIC grant. The goal is to use fusion, the process that powers the Sun, to push uncrewed spacecraft across the solar system far faster and more efficiently than today's rockets allow. StarSplice is built on an unusual idea. It takes the biggest weakness of an older fusion design that has been largely abandoned and turns it into the engine's exhaust which is similar in principle to Sunbird but with a different architecture.
Fusion happens when light atoms are forced together hard enough to merge, releasing large amounts of energy. StarSplice would use deuterium, a heavy form of hydrogen found in ordinary seawater. Getting deuterium atoms to fuse means heating the fuel to roughly 500 million degrees Celsius, more than thirty times hotter than the center of the Sun. At that temperature, the fuel becomes plasma, a gas so hot that its atoms have been stripped into electrically charged particles. No material wall can hold it, so engineers use magnetic fields instead. Charged particles cannot easily cross magnetic field lines, and they spiral along them instead, like beads sliding along a wire.
One of the oldest ways to trap plasma this way is called a magnetic mirror. Picture a straight tube of magnetic field with an extra-strong magnet at each end. As a spiraling particle moves toward either end, the stronger field pushes back on it and turns it around, much as a ball rolling up a steep hill slows, stops, and rolls back down. The particle bounces back and forth between the two ends, trapped. The catch is that particles traveling almost straight along the tube are not turned back, and they escape out the ends. For a fusion power plant, that constant leak wastes energy, and it is the main reason mirrors fell out of favor in the 1980s.
For a rocket, that leak is exactly what you want, because escaping particles are exhaust. StarSplice makes the mirror lopsided on purpose. The magnet at the back is weaker, so most escaping plasma leaves there and is steered out of the engine by a magnetic nozzle, a shaped field that works like the bell of a conventional rocket. The magnet at the front is stronger, so less plasma escapes there. What does escape runs into a direct energy converter, a set of electrically charged grids that slow the incoming particles and collect their energy as electricity, with no turbine or steam cycle in between. The design point calls for exhaust leaving at about 200 kilometers per second, roughly 45 times faster than the best chemical rockets, producing a gentle but continuous push of about 10 newtons, about the weight of a one-kilogram bag of sugar.
The timing matters because the key pieces now exist separately. In 2024, the University of Wisconsin's WHAM experiment held plasma in a mirror using 17-tesla magnets, the strongest field yet applied to a confined plasma, built with new high-temperature superconductors.
Direct energy converters recovered nearly half the energy of escaping plasma in laboratory tests decades ago, and an early demonstration on WHAM was announced in June 2026, though it has not yet been peer reviewed. What no one has studied is how a mirror built unevenly on purpose would behave in the conditions these new machines reach.
The best-known fusion rocket concept today is consequently the Direct Fusion Drive, which makes it excellent for deep space missions. It holds its plasma in a closed, self-contained loop of magnetic field, heats it with a rotating magnetic field, and makes electricity by turning heat into power, much like a conventional power plant. StarSplice differs at each step. Its magnetic field lines are open at both ends. It heats the plasma by firing beams of fast-moving atoms into it. Its electricity comes straight from escaping particles rather than from heat. Much of the physics StarSplice depends on has already been shown to work in the laboratory but has never been applied or proposed as a propulsion concept.
In its current form, fusion returns only one to two percent of the energy the heating beams put in, so the engine would need a separate onboard power source to keep running. Its fuel also gives off substantial neutron radiation, which limits it to uncrewed missions. StarSplice is a concept under study, not a working engine yet.
The proposed nine-month study will answer three questions. Is there a realistic path to fusion supplying more of its own power? Can the whole system be light enough to beat nuclear electric propulsion, where a nuclear reactor powers electric thrusters, today's leading option for deep space? And does a lopsided mirror stay stable, with its escaping energy going where the design needs it? A clear no on any of these would still be valuable, because it would show the field where not to spend the next decade.
Comments