The terminology in the literature surrounding this subject is somewhat confused — the term aerospike originally was used for a (very roughly conically tapering) truncated plug nozzle with some gas injection to form an 'air spike' to help make up for the absence of the tail of the plug. However, frequently, a full-length plug nozzle is now described as being an aerospike.
An engine bell works by confining the sideways flow of the gases, creating a local area of increased pressure with a region of lower pressure "below it". This causes the gases to preferentially flow in the direction of decreasing pressure. By careful design the engine bell grows wider so that the pressure decreases in such a way that by the time the exhaust flow has reached the exit of the bell, it is traveling almost completely rearward, maximizing thrust.
The problem with the conventional approach is that the outside air pressure also contributes to confining the flow of the exhaust gases. At any one altitude, and thus ambient air pressure, the bell can be designed to be nearly "perfect," but that same bell will not be perfect at other pressures, or altitudes. Thus, as a rocket climbs through the atmosphere its efficiency, and thus thrust, changes fairly dramatically, often as much as 30%.
A normal rocket engine uses a large engine bell to direct the jet of exhaust from the engine to the surrounding airflow and maximize its acceleration – and thus the thrust. However, the proper design of the bell varies with external conditions: one that is designed to operate at high altitudes where the air pressure is lower needs to be much larger than one designed for low altitudes. The losses of using the wrong design can be significant. For instance the Space Shuttle engine can generate an exhaust velocity of just over 4,400 m/s in space, but only 3,500 m/s at sea level. Tuning the bell to the average environment in which the engine will operate is an important task in any rocket design.
The aerospike attempts to avoid this problem. Instead of firing the exhaust out of a small hole in the middle of a bell, it is fired along the outside edge of a wedge-shaped protrusion, the "spike". The spike forms one side of a virtual bell, with the other side being formed by the outside air – thus the aero-spike.
The idea behind the aerospike design is that at low altitude the ambient pressure compresses the wake against the nozzle. The recirculation in the base zone of the wedge can then raise the pressure there to near ambient. Since the pressure on top of the engine is ambient, this means that base gives no overall thrust (but it also means that this part of the nozzle doesn't lose thrust by forming a partial vacuum, thus the base part of the nozzle can be ignored at low altitude).
As the spacecraft climbs to higher altitudes, the air pressure holding the exhaust against the spike decreases, allowing the exhaust to expand further as it leaves the engine. This produces an effect like that of a bell that grows larger as air pressure falls, providing altitude compensation. Further, although the base pressure drops, the recirculation zone keeps the pressure on the base up to a fraction of 1 bar, a pressure that is not balanced by the near vacuum on top of the engine; this difference in pressure gives extra thrust at altitude, contributing to the altitude compensating effect.
In theory the aerospike is slightly less efficient than a bell designed for any given fixed altitude, yet it outperforms that same bell at almost all other altitudes. The difference can be considerable, with typical designs claiming over 90% efficiency at all altitudes.
The disadvantages of aerospikes seem to be extra weight for the spike, and increased cooling requirements due to the extra heated area. Further, the larger cooled area can reduce performance below theoretical levels by reducing the pressure against the nozzle. Also, aerospikes work relatively poorly between Mach 1-3, where the airflow around the vehicle has reduced pressure, and this reduces the thrust.
Several versions of the design exist, differentiated by their shape. In the toroidal aerospike the spike is bowl-shaped with the exhaust exiting in a ring around the outer rim. In theory this requires an infinitely long spike for best efficiency, but by blowing a small amount of gas out the center of a shorter truncated spike, something similar can be achieved.
In the linear aerospike the spike consists of a tapered wedge-shaped plate, with exhaust exiting on either side at the "thick" end. This design has the advantage of being stackable, allowing several smaller engines to be placed in a row to make one larger engine while augmenting steering performance with the use of individual engine throttle control.
Three XRS-2200 engines were built during the X-33 program and underwent testing at NASA's Stennis Space Center. The single-engine tests were a success, but the program was halted before the testing for the 2-engine setup could be completed. The XRS-2200 produces 204,420 lbf thrust with an Isp of 339 seconds at sea level, and 266,230 lbf thrust with an Isp of 436.5 seconds in a vacuum.
The RS-2200 larger aerospike engine derived from the XRS-2200. The RS-2200 was to power the VentureStar single-stage-to-orbit vehicle. In the latest design, seven RS-2200s producing 542,000 pounds of thrust each would boost the VentureStar into low earth orbit. The development on the RS-2200 was formally halted in early 2001 when the X-33 program did not receive Space Launch Initiative funding. Lockheed Martin chose to not continue the VentureStar program without any funding support from NASA.
Although the cancelling of the X-33 program was a setback for aerospike engineering, it is not the end of the story. A milestone was achieved when a joint academic/industry team from California State University, Long Beach (CSULB) and Garvey Spacecraft Corporation successfully conducted a flight test of a liquid-propellant powered aerospike engine in the Mojave Desert on September 20 2003. CSULB students had developed their Prospector 2 (P-2) rocket using a 1,000 lbf (4.4 kN) LOX/ethanol aerospike engine. This work on aerospike engines is ongoing; Prospector-10, a ten-chamber aerospike engine, was test-fired June 25, 2008.
Further progress came in March 2004 when two successful tests were carried out at the NASA Dryden Flight Research Centre using small-scale rockets manufactured by Blacksky Corporation, based in Carlsbad, California. The two rockets were solid-fuel powered and fitted with non-truncated toroidal aerospike nozzles. They reached an apogee of 26,000 ft and speeds of about Mach 1.5.
Small-scale aerospike engine development using a hybrid rocket propellant configuration has been ongoing by members of the Reaction Research Society. Another new aerospace research and development group called StoffelCorp Aerospace had recently developed and static tested an aerospike nozzle hybrid rocket configuration with success July 2006. Further aerospike hybrid rocket motor tests were scheduled for 2007.