Since the 1950s, nuclear rockets were developed primarily at Los Alamos National Laboratory to produce faster methods for space travel. (Bussard and DeLauer, 1958; Dewar, 1974; Borowski, 1987; Dewar, 2007). These technologies utilize nuclear designs that transfer heat from a sealed core to liquid hydrogen expanders or thermionic converters in a conventional manner. Beginning in the 1980s, a more efficient nuclear energy transfer design emerged for rocketry, (Haslett, 1995; Lieberman, 1992), one which exposes the core for direct nuclear fragment thrust once the rocket is well outside the Earth’s atmosphere. From FY11 through FY14, the NASA Institute for Advanced Concepts studied the Fission Fragment Rocket Engine (FFRE). (Werka et al., 2012; Chapline, 1988; Chapline et al., 1988; Chapline and Matsuda, 1991). The FFRE would directly convert the momentum of fission fragments into spacecraft momentum at extremely high specific impulse (I _SP). The I _SP is a measure of how efficiently an engine uses fuel to create thrust. For rocketry, I _SP is defined as the thrust integrated over time per unit weight (on Earth) of the propellant. (Benson, 2008; Sutton and Biblarz, 2016). The I _SP is given by Eq. 1 I S P = v e g 0 (1) where g 0 is the gravitational acceleration constant on Earth and v e is the average exhaust velocity.

The FFRE combined the existing technologies of low-density dust trapped electrostatically (dusty plasmas) with high field magnets for fission fragment control. By designing the nuclear core to permit sufficient mean free path for the escape of fission fragments, the study showed the FFRE could convert nuclear power to thrust directly and efficiently at a delivered I _SP of 527,000 s. The study showed further that, without increasing the reactor power, adding a neutral gas to the fission fragment beam significantly increased the FFRE thrust. This frictional interaction of gas and beam resulted in an engine that would continuously produce nearly 4.5 kN of thrust at a delivered impulse of 32,000 s, thereby reducing the (then expected) round trip mission to Mars from 3 years to 260 days. These studies concluded that the engine and spacecraft were within current technological capabilities, could be built and launched, though the FFRE engines would be very large. New materials, and recent advances in technology, promise to make FFREs much more compact and reliable.

For manned transit to Mars, speed is of the essence to minimize radiation exposure and maximize human health, as discussed in other articles in this volume. Nuclear fuel, with several million times the energy density of any chemical fuel, holds the promise of fast transit to Mars. Typically, two types of nuclear propulsion are considered for space flight: thermal propulsion and electric propulsion.

In thermal propulsion, a coolant that prevents reactor melt-down doubles as a propellant. This gas is heated by direct interaction with the fuel, which limits the gas heating to a temperature below 3000 K, where typical fission fuel melts. A metric of the efficiency of a rocket engine is the specific impulse or I _SP. The I _SP is thrust (force) of the engine divided by the weight flow of the propellent. The higher the I _SP, the more efficient the engine. In nuclear thermal propulsion I _SP of approximately 900 s can be achieved, compared to perhaps half that for the best chemical engines. While certainly an improvement, with nuclear energy’s enormous energy density, conventional nuclear thermal propulsion could benefit from technological innovation.

In electric propulsion, the nuclear reactor is configured to produce electricity, and that electricity is used to drive a conventional ion or plasma source that provides propulsion. These sources tend to have high I _SP, but low thrust. Thrust, approximately the velocity of ejected propellent times mass flow, is required to lift a rocket out of a gravity well. For a mission to Mars, a rocket engine would ideally have the capacity to produce high thrust (for leaving the Earth, Moon, or Mars) and then high I _SP (high efficiency) for interplanetary transit. As an analogy, this is similar to an automobile starting with a low gear and high torque, finally traveling at high speed in a high gear. In a reactor producing electricity, Carnot efficiency requires a large difference in output temperature compared to input temperature. In a vacuum, space craft have difficulty rejecting heat, requiring large radiators. Increasing input temperature is again limited by the melting temperature of the nuclear fuel.

For both of these technologies, nuclear electric propulsion or nuclear thermal propulsion, performance enhancement can be achieved through dealing with the fuel heating problem. If we could heat the propellent to a higher temperature than the melting point of the fuel, then we could achieve higher thrust. If we could operate the nuclear reactor at a temperature well above the melting point of the nuclear fuel, the Carnot efficiency could also be improved.

Though not currently applied to propulsion, the NASA Kilopower space-reactor design uses a 6-inch diameter core of enriched uranium surrounded by a beryllium-oxide reflector. A boron-carbide rod is used as a reactor switch. The design should provide between 1 and 10 kW of power. These surface power technologies will also guide advancements in designing propulsion systems. (NASA, 2017; NASA, 2018; NASA, 2022). However, to achieve a sufficient I _SP, any FFRE designs would require a much higher power output. FFRE designs would likely be operated as a subcritical assembly.

Fission fragment nuclear propulsion was proposed by George Chapline of Lawrence Livermore National Laboratory in the late 1980s. (Chapline, 1988; Chapline et al., 1988; Chapline and Matsuda, 1991). The propulsion method depended on the fission fragments of a nuclear reaction partially escaping the fissile fuel, mitigating the heating of the fuel. When a fissile material undergoes fission, two daughter atoms are produced as fragments of the event. These atoms carry a total energy of 160 MeV and can travel at velocities up to 5% of the speed of light. If the fission occurs in a very thin layer of nuclear fuel (a few microns thick), some of the fission fragments can escape without collision, minimizing fuel heating. With a magnetic field used to direct these fragments to the exhaust direction, extremely high I _SP could be achieved, perhaps on the order of 10⁶ s. This would be two orders of magnitude larger than an electric ion source and more than three orders greater than chemical or nuclear thermal propulsion. On long transit missions this increased I _SP, and the decrease in mass of the rocket as its nuclear fragments are ejected, would result in an extremely high final rocket velocity. Chapline envisioned the thin fissile fuel coating being placed on carbon fibers, the fibers moving at high speed through a neutron moderator region with many other coated fibers, having sufficient fissile mass to sustain the reaction, obtaining reactor criticality. The fibers would then exit the neutron moderator to radiatively cool in the vacuum of space. Even though many fission fragments would escape the fuel and eject at high speed, the carbon fibers would still heat significantly. However, the carbon fibers would still contribute to enhancing the thrust.

To simulate a 14-MeV neutron source incident on a layer of uranium oxide, a nuclear transport code (MCNP6.2) was employed. (Werner et al., 2017). The simulation tracked any heavy ions generated to estimate the percentage that escaped as a function of the thickness of the uranium oxide layer. The total amount of fission fragments created increases with increasing UO₂ thickness. This amount only begins to increase by orders of magnitude for layers thicker than 1 mm. The fission fragment population remaining in the UO₂ increases with layer thickness. Since a larger fraction of the fission fragments remains in the UO₂ for thicker layers, a smaller percentage of the total created fragments escapes the layer. The results of the simulation, illustrated in Figure 1, confirm that nearly 100% of heavy ions can escape for layer thicknesses below a few microns. Although thicker uranium oxide layers generate more fission fragments, almost all heavy ions remain trapped inside the fissile material. Therefore, using fissile layers thicker than approximately 10 microns does not yield any benefits to the I _SP.

To avoid this heating, NASA initiated the study of the Fission Fragment Rocket Engine (FFRE) concept in 2010. (Clark and Sheldon, 2005; Werka et al., 2012; Clark et al., 2013). The FFRE would directly convert the momentum of fission fragments into spacecraft momentum at extremely high I _SP. The FFRE combined the existing technologies of low-density dust trapped electrostatically (dusty plasmas) with high field magnets for fission fragment control. The fissile fuel fragments and dust would be micron-sized, allowing for the fission fragments to escape the fuel particle without depositing significant energy into the fuel particle or heating it. By designing the nuclear core’s uranium dust density to permit a sufficient mean free path for the escape of fission fragments, while simultaneously maintaining density sufficient to maintain criticality, the study showed the FFRE could convert nuclear power to thrust directly and efficiently at a delivered I _SP of 527,000 s.

Without increasing the reactor power, adding a neutral gas to the fission fragment beam significantly increased the FFRE thrust. This frictional interaction of gas and beam resulted in an engine that would continuously produce nearly 4.5 kN of thrust at a delivered impulse of 32,000 s, thereby reducing the (then expected) round trip mission to Mars from 3 years to 260 days. It importantly would allow the heating of the gas propellent to a very high temperature, improving thermodynamic efficiency without melting the uranium oxide fuel. The fuel dust would still heat from the gamma radiation in the reactor, however the small size of the fuel dust would mean that the particles had a larger surface-area-to-volume ratio compared to large particles, enhancing radiative cooling of the fuel.

These studies concluded that the engine and spacecraft were within current technological capabilities, could be built and launched, however the FFREs would be very large. To sustain a nuclear reaction, a neutron produced by fission in the fuel should be moderated and available to initiate another fission event in order to “go critical.” There also needs to be a sufficient amount of fuel mass to continue the reaction. In the FFRE concept, the fuel density needs to be low enough for fission fragments to get out. The design constraint is to design a reactor that keeps neutrons in, and allows fission fragments out, and fundamentally this means a low-density fuel in a large volume.

In an early conceptual design, superconducting magnets were used adjacent to the nuclear reaction. With one magnet having a radius of over 5 m, the coils would have a combined mass of over 40 metric tons. Either a moderator or neutron reflector would be needed to keep neutrons in the reactor, maintaining criticality. This would add over 50 metric tons and the total engine mass would be over 113 metric tons. (Clark et al., 2013). In this design, the extra mass would not be expended as the rocket accelerated, substantially limiting the final velocity of the rocket.

Though there are a number of other very practical issues to consider for the successful deployment of a dusty plasma FFRE, perhaps most importantly, no one has ever built or tested a dusty plasma reactor on Earth, let alone in space.

Fast transit to Mars poses several challenges to rocket design due to overall mass, specific impulse, thrust, and thermal management. It may be possible to circumvent several of these obstacles by using subcritical assemblies, energy efficient neutron generation, and compact, high-field magnets to generate and steer fission fragments. Integrating these innovations into FFRE designs may provide increased specific impulse and controlled thrust while operating as a subcritical assembly. These new technologies and improvements may be integrated into a revolutionary nuclear rocket technology and dramatically reduce transit time to Mars.