In his paper describing the concept, Zubrin considers using a Nuclear Salt Water Rocket for a round trip mission to Titan, Saturn's largest moon. The NSWR would be fueled by 20% enriched uranium in the chemical form of a soluble salt (uranium tetra-bromide) dissolved in ordinary water at about the same atom number concentration as the salt in sea water.
Fissionable isotopes in such concentrations can easily produce great heat from fission reactions or even a nuclear explosion. An uninterrupted volume of this liquid massing a few dozen kilograms would reach critical mass, massively fission in a sustained chain reaction, and explode. In Zubrin's scheme 41,000 kilograms (41 tonnes) of the salt water fuel are stored in a neutron-absorbing fuel tank. The fuel tank would be made from long tubes of boron carbonate, a strong structural material that strongly absorbs thermal neutrons, preventing the fission chain reaction that would otherwise occur in the fuel. The liquid fuel is pumped from the storage tank into a absorber-free cylindrical reaction chamber which allows buildup of neutron flux to the critical point where sustained nuclear fission can occur.
In a nuclear rocket the reaction chamber presents a severe materials problem because no conceivable mechanical structure could sustain the force of a nuclear explosion. However, Zubrin uses a very clever trick. He has used a simplified model to show that the distribution of fission-inducing thermal neutrons in the reaction chamber depends critically on the velocity of the liquid fuel as it passes through the reaction chamber. This dependence occurs because the moving salt water fuel is also the medium in which the neutrons are slowed. If the liquid is at rest, the maximum flux occurs at the center of the cylinder, but if the moderating fuel liquid is in motion, the point of maximum flux is skewed downstream and also rises to a much higher maximum. If the right fuel velocity is chosen, the thermal neutron flux (and therefore the site of maximum fission energy release) can be made to peak very sharply just outside the exit end of the cylindrical reaction chamber.
In other words, one can produce a continuous controlled nuclear explosion in the region just behind the nuclear rocket. At this point the water of the fuel liquid flashes to very high temperature steam, expelling reaction mass with an estimated exhaust velocity of 66,000 meters per second (as compared with perhaps 4,500 m/s for a chemical rocket). The NSWR engine is calculated to produce a thrust of almost 3 million pounds (1.3 x 107 N) and to have a power output of 427 gigawatts. With this kind of performance, the mission to Titan could be launched from low earth orbit with an acceleration of almost 4 g's and could, in principle, be carried out with low launch mass, low cost and high efficiency.
Zubrin also considers how a NSWR might be used in a more ambitious 120 year one-way probe mission to Alpha Centauri. He envisions a 300 tonne spacecraft carrying 2700 tonnes of salt water fuel containing 90% enriched uranium. This highly enriched fuel would be burned in a high efficiency engine to produce an exhaust velocity of 4,700,000 m/s, permitting the spacecraft to achieve a velocity that is 3.63 % of the velocity of light. He proposes to use most of the fuel for acceleration and to use a magnetic sail (see Analog, May-'92) for deceleration by creating drag against the interstellar medium.
What Zubrin has described, therefore, is a high-energy space propulsion technology suitable for deep space and interstellar missions that could be implemented with fairly modest extensions of current technology. Moreover, the end of the cold war has left in its wake considerable stockpiles of fissionable materials (239Pu and highly enriched 235U) from decommissioned nuclear weapons that can be regarded as a source of cheap fuel for such projects. Zubrin also points out that, despite the highly radioactive exhaust of the NSWR, the engine itself need not be radioactive to any significant degree. The fuel has only low-level alpha activity, the fission products from the consumed fuel are vented into space, and the induced activity from the large neutron flux produced by the fission burning can be minimized by constructing the engine from such low activation materials as graphite and silicon carbide. Once the engine is turned off, therefore, there should be no significant radioactive inventory present to endanger the crew of a manned mission.
The highly radioactive exhaust, of course, constitutes a major disadvantage for the NSWR scheme. The prospect of contaminating space with radioactive waste is certain to draw strong opposition from the same environmental and anti-nuclear groups that have opposed the use of nuclear power sources in NASA's deep space missions. Zubrin argues, however, that the NSWR's exhaust velocity of 66 km/sec far exceeds the escape velocity of any planet, and that as long as the exhaust vector does not intersect the Earth "the amount of contaminant reaching the Earth could be insignificant" even in an NSWR launch from low earth orbit. In fact, since the atoms of exhaust gas have sufficient velocity that they are not bound by the Sun's gravity well, the expelled exhaust will dissipate rapidly and will soon leave the Solar System altogether. Zubrin also points out that since the NSWR is not a weapon or a bomb, its testing and use does not violate the 1968 Test Ban Treaty. Therefore, unlike the Orion scheme, its use as a space propulsion system is legal.
There would, of course, be some very demanding technical challenges in designing a safe and reliable NSWR. The extremely high exhaust temperature and velocity of the device present a particular challenge in designing an exhaust nozzle for the NSWR that will not be severely eroded during a brief period. Zubrin suggests that a continuous flow of normal (unsalted) water along the surface of the reaction chamber and nozzle could provide cooling and extra reaction mass, but this remains to be demonstrated. A full design would also have to consider possible failure modes, including the possibility of a fuel pump failure that could cause a fuel detonation within rather than behind the reaction chamber. These appear to be solvable problems, but they would have to be addressed.
In summary, the NSWR appears to be a radical but feasible solution to the problem of mounting an interstellar mission with essentially existing technology. An unmanned Alpha Centauri probe of the type that Zubrin suggests could be built starting today and at a cost that I would guess would be much smaller than the growing price tag of NASA's troubled Space Station Freedom project. The anticipated mission time of 120 years is a long time. But the sooner we start, the sooner we (or our descendants) will get a closeup look at our neighboring star systems.
I had not heard of this concept; I'd like to see more on it. Do you have links?
A "conventional" nuclear-thermal rocket (NERVA or KIWI) simply uses a solid nuclear reactor which is cooled by hydrogen. It can get 800 seconds or possibly higher. As I mentioned, thrust-to-weight is a problem.
One question: if this concept relies on the fluid velocity to prevent an "explosion", how do you start it up...or shut it down?...
--Boris