|Fig. 1: The Orion 10-m Test Design.  (Courtesy of NASA)|
Since nuclear reactions provide an enormously large energy-to-mass ratio compared to chemical reactions (e.g. a typical fission event releases 200 MeV compared to a few eV for a chemical bond) it is no surprise that nuclear energy is a prime candidate for realistic high-performance spacecraft engines. One basic design is the thermal nuclear rocket, for which actual working engines were developed by NASA in the NERVA project of the 60s and 70s. In this rocket, standard liquid hydrogen fuel is heated using a nuclear reactor to high temperatures and expelled through a thrust nozzle; the high temperatures achieved through nuclear processes can allow for more than twice the effective exhaust velocity of normal engines . The downside to this design, along with similar internal nuclear reaction designs, is that the exhaust velocity achieved is limited by the temperatures that can be withstood by the engine components, and the performance (specific impulse) of the rocket increases as the square root of the exhaust temperature. 
In 1947, Stanislaw Ulam created a radical new design for a nuclear-propelled rocket that would attempt to circumnavigate this problem: the external nuclear-pulse propulsion engine. He proposed ejecting a series of nuclear explosives encased in reaction mass from the back of the rocket; the explosion would eject a portion of the reaction mass at high velocity into a cushioned pusher-plate at the back of the ship. While this may seem incredibly dangerous, the interaction between the blast wave and the plate lasts for around a millisecond, yielding a useful transfer of momentum with very little heat transfer. Simple protection methods can prevent even the small amount of ablation due to this interaction from occurring . The advantage of the high-velocity thrust afforded by a nuclear explosion is that the nuclear-pulse engine can achieve very high thrust force along with very high efficiency, two qualities which were impossible to achieve together in chemical rockets. Other physicists were suitably impressed with this design, and Ted Taylor along with Freeman Dyson initiated a NASA-funded design study at General Atomics in 1958. The project was given the name "Orion."
|Fig. 2: Trhust vs. specific impulse (Isp) for different size Orion nuclear-pulse models.  (Courtesy of Nasa)|
I will primarily discuss one specific Orion design which was thoroughly analyzed theoretically in 1963 under Air Force contract. This vehicle was 91,000 kg in weight and 10m in diameter; this is actually a great deal smaller than previous designs.  Surprisingly, much better efficiencies are achieved in the case of larger vehicles due to the greater efficiency of larger explosives. However, NASA decided on this design as an economic compromise due to the cost of larger vehicles.
Like all rockets, the vehicle consists of a crew/payload section on top of the thrust portion of the rocket. The shaped nuclear explosive charges are carried in racks both internally and externally; this particular design has room for 900 units in the internal racks, with the ability to carry more as needed externally. The pusher plate is a thin steel disk, with thickness varying with radius to minimize bending stresses during operation. A hole in the plate allows for the distribution of anti-ablative oil on the plate between pulses. Between the plate and the payload section are two shock-absorber systems: the first, connected to the plate directly, is a series of gas-filled tori maintained at 100 psi, similar in function to automobile tires. These are connected to a more rigid secondary spring system. Directly above this secondary shock-absorber are the magazine racks. Small chemical thrusters are used to maintain precise control of the ship in flight. Shielding is used on the crew section to protect from radiation from the nuclear explosions. 
|Fig. 3: Shaped nuclear pulse unit for the 10 m Orion design. Larger Orion models utilize more powerful, efficient devices.  (Courtesy of NASA)|
In the standard operating mode for this Orion prototype, approximately one pulse unit per second is ejected through the center of the pusher plate, and detonation would then occur. Oil, sprayed onto the surface of the plate after every pulse, provides excellent protection against ablation, but the plate still experiences a maximum acceleration on the order of 50,000 g.  The two shock absorber systems are designed to reduce this acceleration to a few g's in the crew section, and these shock absorbers oscillate multiple times per pulse cycle.
As shown in this diagram from , this particular nuclear-pulse engine generates about 3.5 × 106 N of thrust with a specific impulse of about 2000 s. Note that in rocketry specific impulse is defined as change in momentum per unit weight-on-earth of propellant, giving it units of seconds; specific impulse is a way of measuring average exhaust speed, and therefore the efficiency of an engine with respect to reaction mass. The specific impulse of the particular nuclear pulse engine under consideration is on the same order of magnitude as an ion thruster, but with much higher thrust (so a much greater acceleration); on the other hand the thrust is comparable to chemical rockets, but the specific impulse is much greater.
The individual pulse units are actually shaped charges, as seen below. The radiation case is designed to reflect energy into the channel filler, which expands rapidly and propels the dense plug of propellant material into the pusher plate.
© Ben Stetler. The author grants permission to copy, distribute and display this work in unaltered form, with attribution to the author, for noncommercial purposes only. All other rights, including commercial rights, are reserved to the author.
 J. C. Nance, "Nuclear Pulse Propulsion," IEEE Trans. on Nuclear Science 12, 177 (1965) [Reprinted as Ann. N.Y. Acad. Sci. 140, 396 (1966)].
 "Nuclear Pulse Space Vehicle Study, Vol III," Report on NASA Contract NAS 8-11053, General Atomics, GA-5009, 19 Sep 64.
 "Nuclear Pulse Vehicle Study Condensed Summary Report," NASA Contract NAS 8-11053 General Atomics, GA-4891, 14 Jan 64. (see )
 W. H. Robbins and H. B. Finger, H.B., "An Historical Perspective of the NERVA Nuclear Rocket Engine Technology Program", NASA Contractor Report 187154, AIAA-91-3451, July 1991.