* This chapter concludes the discussion of advanced space rocket propulsion systems with discussions of electrostatic ion drives, other related electric propulsion schemes, and various advanced propulsion concepts now in consideration. A footnote is provided with a history of space nuclear power, since it is relevant to the use of space nuclear power for propulsion systems.
* Another advanced propulsion concept that has been around for about as long as NTR engines are "electrostatic ion propulsion (EIP)" drives. The first EIP drive was built in 1958 by NASA. The first flight test was in 1964, with the NASA "Space Electric Rocket Test (SERT)" suborbital flight, and was followed by further test flights -- but the technology only reached full operational status in the 1990s.
The EIP drive is one of several different types of electric rocket engines, and in fact is not the first type to be used in space. However, it is conceptually one of the simplest and best-known. The basic idea is to accelerate ions of a relatively heavy propellant, such as xenon, to high velocities, through grids charged to high voltages. The thrust is very low, but the efficiency is very high, with a specific impulse over ten times that of LOX-RP propulsion.
Early experimental EIP drives used mercury or cesium as a propellant. They were troublesome since handling such toxic propellants was tricky, and the propellants had to be vaporized for use, involving energy loss through heat of vaporization. Although xenon is rare, it has almost ideal properties for EIP drives: it is normally a gas, has a high atomic mass and low ionization potential, is chemically inert, and is easy to handle.
Hughes Space (now part of Boeing) was the leader in developing the technology, which the company calls the "Xenon Ion Propulsion System (XIPS)". Hughes had experimented with EIP drives since the 1960s, and judged xenon to be the propellant of choice in 1984. A formal program to develop operational EIP drives was begun in 1992.
The company attracted particular attention with the EIP drive developed for the NASA "Deep Space 1 (DS1)" probe, which was launched on 24 October 1998. The spacecraft was one of the first launched under the NASA "New Millennium" program. Although DS1 was to visit asteroids and comets, the goal of New Millennium is to develop and demonstrate advanced spacecraft technologies, and the probe incorporated twelve such advanced technologies, most significantly the EIP drive.
The EIP drive was powered by the spacecraft's photovoltaic arrays, and so was referred to as a "solar electric propulsion (SEP)" drive. DS1 was the first spacecraft to use an electric rocket engine of any type for its main propulsion system. DS1 weighed 490 kilograms, was 2.5 meters long, and was 1.7 meters wide. When its two solar panels were deployed, they spanned 11.8 meters. DS1's EIP drive was named the "NASA SEP Technology Application Readiness (NSTAR)". NSTAR weighed 8 kilograms, had a diameter of 30 centimeters, and used a supply of 81.5 kilograms of xenon propellant.
Xenon atoms were fed from a propellant tank into an ionizing chamber, where an electron emitter ionized them, producing positive xenon ions. These xenon ions were then accelerated by two molybdenum grids with a potential of 1,300 volts between them, driving the ions out the exhaust at over 30 kilometers per second. Electrons were injected into the external exhaust flow to neutralize the positive ions; this prevented the spacecraft from building up a negative charge, which would have attracted exhaust ions back to the grid and possibly disrupted the operation of spacecraft instruments.
An EIP drive can be throttled either by adjusting the grid voltage or adjusting the propellant flow. NSTAR adjusted the propellant flow. EIP engine efficiency is reduced by the energy needed to ionize and then neutralize the propellant, and leakage of unionized propellant. NSTAR's conversion efficiency of electric power to thrust was 63%, and the engine could be throttled over a range of 2 to 9.4 grams of thrust. Design life of the drive was 8,000 hours. Development of NSTAR was begun in 1993 by NASA's Jet Propulsion Laboratory (JPL), which runs the New Millennium program. Hughes built most of NSTAR, though Moog INC built the xenon fuel system.
* The DS1 mission went smoothly, though initial attempts to fire up the NSTAR EIP drive encountered difficulties. The first test was to last 17 hours, but the drive shut down after 4.5 minutes.
The two ion-accelerating grids were only about 0.6 millimeter apart, and this narrow spacing could result in shorts and arc-overs, possibly due to molybdenum flaking off the grid or other contaminants. When a short or arc-over was detected, the power system "recycled" -- that is, it shut down for a second and then restarted, which usually cleared the fault. However, the EIP drive exceeded the limit on recycles and mission engineers were puzzled. They were not particularly alarmed, because such behavior was familiar from ground tests of the engine, and they knew that recycles were common during initial operation until the engine burned away contaminants. NSTAR eventually operated for much longer than its specified design life.
DS1's advanced solar arrays used reflectors and lenses to concentrate sunlight on photovoltaic cells. The panels could generate 2.6 kilowatts for the NSTAR engine and the spacecraft's other systems at the distance of the Earth from the Sun, though power fell off as the probe moved farther into space. The NASA Dawn asteroid probe, launched in 2007, used three EIP drives based on DS1 technology.
The Japanese Hayabusa asteroid sample-return probe, launched in 2003, used four EIP drives; they were of Japanese design, but may have leveraged off US expertise, the Japanese often doing so in space activities. The Hayabusa mission didn't go well, but the EIP drives were not really part of its problems. A largely similar Hayabusa 2 was launched in 2014, which proved much more successful.
* Hughes developed two versions of smaller EIP drives, which have been used operationally on communications satellites ("comsats") as station-keeping thrusters. Comsats are launched into "geostationary" orbit, at an altitude of about 36,000 kilometers. At this altitude, the satellite takes 24 hours to orbit the Earth. This matches the Earth's rate of rotation, and so the satellite hangs in the same position above the Earth. This is useful for a comsat, since ground-based antennas can be focused on the satellite and then left in place. The orbit has to be directly over the equator, or the satellite will move north and south over the course of the day. Since the Sun is not in the plane of the Earth's equator, its gravitational pull tends to nudge a geostationary satellite's orbit off its equatorial position. The Sun's influence is very small, and so even a weak thruster is powerful enough to keep the satellite on station.
The first operational EIP drives were flown on the "PanAmSat PAS-5" comsat in 1997, which was a Hughes (now Boeing) HS-601HP spacecraft fitted with four EIP drive thrusters for station-keeping. Each thruster was 13 centimeters in diameter and generated 1.3 grams of thrust. Design lifetime for the thruster system was 12 to 15 years. Only two thrusters were actually used, pointing north and south to keep the satellite in its orbit plane, with the other two reserved as backups. Each thruster operated for about five hours a day.
Hughes also developed a more powerful EIP drive thruster for use with the larger HS-702 comsat. The bigger thruster has a diameter of 25 centimeters and produces 16.8 grams of thrust. As with the HS-601HP, four thrusters are fitted, with two being redundant. The greater thrust of the larger thruster means that a thruster only needs to be fired for about a half hour a day to keep the satellite on station. The product line is now associated with Boeing, which bought up the Hughes satellite arm.
Comsats with electric drive thrusters are now in widespread service. EIP drive thrusters cut fuel requirements from those of traditional hydrazine chemical thrusters by as much as 90%. This results in weight savings of hundreds of kilograms, and millions of dollars saved in payload launch costs.
* Work on next-generation EIP drives is in progress. Operational lifetime is a problem. EIP drives suffer from ionic and atomic erosion of the grids, which eventually wears the grids out. One engineer compares it to "sandblasting on the atomic scale." Advanced materials with longer lifetimes are being evaluated. EIP drives are now being developed with much greater lifetimes and lower costs, and in sizes ranging from smaller to larger than current EIP drives. NASA has worked on EIP drives that are an order of magnitude more powerful than the DS1 NSTAR and have two to three times the fuel economy.
Scaling up the electrical grid for an EIP drives runs into some physical obstacles, and so a large spacecraft would use multiple EIP drives to obtain more thrust. For such big spacecraft, the solar panels required to provide power for the EIP drives would be prohibitively large, and so such a spacecraft would likely be powered by a nuclear reactor.
NASA and Northrop Grumman did perform research on a program named "Prometheus" to develop a nuclear-powered EIP spacecraft. The main focus of the Prometheus effort was the "Jupiter Icy Moons Orbiter (JIMO)", which would have been by far the biggest space probe ever flown, with a launch mass of 20 tonnes and a deployed length of 30 meters. The JIMO design featured a nuclear reactor with an output power of a few hundred kilowatts at its tip, linked to a spacecraft bus by a long boom flanked by heat radiators and with twin EIP engines at the end.
Liquid metal and gas cooled reactor designs were considered for JIMO. Although JIMO was to be an operational spacecraft, its primary rationale was as a demonstrator for a reactor for crewed space missions that would be ten times more powerful. The JIMO program was too ambitious, and was scaled back to an investigation.
BACK_TO_TOP* The European Space Agency (ESA) has flown a demonstrator spacecraft, the first "Small Mission For Advanced Research In Technology (SMART-1)", to test their own SEP drive. SMART-1 was a Moon orbiter, with a small payload of experimental instruments to perform lunar measurements to obtain some science value from the mission, and was launched as a secondary payload on an Ariane 5 booster in September 2003.
The SMART-1 SEP system used a "Hall effect drive", with magnetic fields, not electrostatic fields, accelerated heavy ions. There are a number of different schemes for Hall effect drives, but one simple configuration consists of a cylinder ringed by one pole of a magnet, and with the other pole configured as a rod running the center of the cylinder. This configuration sets up a magnetic field running radially from the center rod to the ring.
A propellant that can support an electric arc discharge, usually xenon, is injected into the inlet of the engine. A positive electrical anode is placed in the cylinder before the magnetic ring, and a negative cathode emitter is placed after the magnetic ring, possibly at the output. The potential between the anode and cathode ionizes the xenon. Since electrical charges move at a right angle to a magnetic field, the radial magnetic field causes the electrons and ions to circle around the center pole. An interaction between the electric and magnetic fields known as the "Hall effect" accelerates the xenon ions out the exhaust.
Hall effect drives have the advantage that they do not require grids, and so in principle can be more easily scaled up than EIP drives. They provide more thrust than EIP drives, but they are not quite as efficient.
The SMART-1 SEP system was based on a French SNECMA PPS-1350 Hall effect thruster, built in collaboration with Russian propulsion organizations. The engine had 7 grams of thrust and slowly nudged SMART-1's Earth orbit outward in a spiral to finally be captured by the Moon's gravity, with the probe finally going into orbit around the Moon in November 2004. An improved Hall effect engine was used on the ESA Bepi-Colombo Mercury probe, launched in 2018. The PPS-1350 has also been integrated into satellites for station-keeping and orbital adjustment.
Both the US and the USSR experimented with Hall effect drives in the 1960s. The US gave up on them in favor of EIP drives, but the Soviets continued their efforts, and have flown at least a hundred of them on various spacecraft. In the last few years, American researchers have evaluated Russian Hall effect drives and found them impressive, and NASA is now working with US aerospace firms to build Hall effect drives.
In 2002, the NASA Glenn Research Center in Ohio announced that they had developed and ground-tested a new Hall-effect thruster, the "NASA-457M", claimed to be ten times more powerful than any other Hall-effect thruster built to that time, capable of providing over 300 grams of thrust.
In 2005, Aerojet announced sale of a Hall effect thruster, to be used for station keeping and orbital adjustment. That led to the "BPT-2000" and "BPT-4000" thrusters, with Aerojet thrusters being flown on comsats since 2010.
BACK_TO_TOP* In addition to the EIP and Hall-effect drives, simpler types of electric rocket engines were in use decades before EIP drives were employed operationally.
"Electrothermal rocket (ETR)" drives operate on much the same principle as an NTR or STR, except that the propellant is heated electrically, instead of by a nuclear reactor or a solar mirror. There are two types of ETR drives, "arcjets" and "resistojets":
There is also an "electromagnetic rocket (EMR)" drive that is similar to an ETR. It is fueled by a solid cylinder of teflon plastic, backed by a feed spring mechanism, with the output end of the block vaporized by a high-intensity current pulse that also sets up an electromagnetic field to accelerate the vaporized teflon out the nozzle. The thrust can be very precisely controlled by varying the magnitude of the current pulse or its repetition rate, and the specific impulse is excellent, from about 2.5 to 5.5 times that of LOX-RP propulsion.
This type of drive is sometimes called a "pulsed plasma thruster (PPT)". The EMR drive has been around a long time, having been first used on the Soviet Zond 2 space probe in 1964. The PPT went out of fashion as satellites got bigger, but now that "smallsats" are making a comeback, interest in PPTs is reviving. NASA's "Earth Observation 1 (EO-1)" satellite featured a 4.95 kilogram PPT, with enough teflon fuel for 30 days of operations.
These types of engines have been used as thrusters, and they offer improved fuel efficiencies in comparison to traditional hydrazine thrusters. Scaling them up to any size seems to be out of the question.
BACK_TO_TOP* Several other types of advanced rocket engines are now in laboratory development, but have never been flown on a space mission. They include the "magnetoplasmodynamic" drive; the "pulsed inductive" drive; and the "variable specific impulse magnetoplasmodynamic rocket (VASIMR)" drive.
In a magnetoplasmodynamic drive, sometimes also called a "Lorenz force accelerator", the propellant is accelerated by magnetic, rather than electric, fields. The engine consists of a thrust chamber with walls that act as an anode and a central rod that acts as a cathode. The propellant, which can be argon, lithium, or hydrogen, in increasing order of efficiency, is ionized, causing a very strong current to flow radially between the anode wall and the central cathode. While the current flow is provided by equal numbers of electrons and positively-charged ions, the ions are much heavier than the electrons and so move much more slowly.
A moving current sets up a magnetic field acting at right angles whose magnitude is proportional to the rate of current flow, and so the fast-moving electron current sets up a magnetic field directed in concentric rings around the cathode. A magnetic field in turn accelerates charged particles at right angles to itself, and so the positive ions are driven out the exhaust nozzle.
* Conceptually, a pulsed inductive drive looks like a flat hoop coil with a fat spike in the center, with the coil connected to a bank of big capacitors. A puff of propellant, usually argon though many other propellants are possible, is injected inside the hoop, and then the capacitors are discharged into the coil. This sets up an intense magnetic field that ionizes the propellant, and the electric fields that are set up push the ions out the hoop. Specific impulse should be ten to twenty times that of LOX-RP propulsion.
Pulsed inductive drives do not have electrodes, which tend to be worn down by ion and electron bombardment, and engine thrust can be scaled up by increasing the pulse rate, which is on the order of several hundred times a second. TRW has been working on the concept using company funds, but no pulsed inductive drive has been flown in space yet.
* The VASIMR drive is one of the most exotic of the advanced propulsion concepts. It is the brainchild of Franklin Chang-Diaz, a Costa Rican-born American plasma physicist who was also a NASA space shuttle astronaut, performing seven space flights -- a record for number of flights at the time of his retirement from the agency. It uses hydrogen or argon as a propellant, first ionizing it with radio-frequency (RF) energy, and then injecting it into a thrust chamber where oscillating magnetic fields and RF energy heat it to millions of degrees Celsius. A magnetic choke controls the flow of the hot plasma to the exhaust nozzle.
If the magnetic choke is constricted, the flow of plasma is small, but the temperature remains high. This gives low thrust but extremely high efficiency, possibly a hundred times that of LOX-RP propulsion, useful for interplanetary cruise. If the magnetic choke is opened up, the flow of plasma is high, but the temperature is low. This gives high thrust and lower efficiency, about ten times that of LOX-RP propulsion, useful for initial boost out of planetary orbit.
Such dual-mode operation is extremely attractive for a crewed Mars mission, since attempting to boost out of low Earth orbit with a low-thrust engine would leave the Mars vehicle in the Earth's radiation belts for a long period of time, putting the crew at risk. The VASIMR engine would allow the Mars vehicle to cross the radiation belts in a short time in the high-thrust mode, and then cruise to Mars in the low-thrust mode, with a total one-way trip time of eight weeks. Work has been conducted on VASIMR at the Massachusetts Institute of Technology, NASA JPL, and in particular at the Ad Astra company of Houston, Texas, set up by Chang-Diaz. So far, there are no specific plans to fly a VASIMR engine.
BACK_TO_TOP* Although nuclear power has not been used to date for space propulsion systems, it has been used as a relatively small-scale power source for spacecraft systems. The history of space nuclear power in general illustrates some of the technical and political issues confronting space nuclear propulsion.
In the 1960s, alongside ambitious nuclear propulsion schemes such as NERVA, the US developed "radioisotope thermoelectric generators (RTG)", in which the heat from decaying radioactive isotopes generated electricity from thermoelectric junctions in the RTG module. The junctions consisted simply of two dissimilar metals joined together that produced an electric current when heated. The first atomic-powered spacecraft, the US Navy's Transit 4A navigation satellite, was launched in 1961 with an RTG power system. The US successfully launched three more RTG-powered satellites, but the fifth, Transit 5BN-3, failed to make orbit when it was launched in April 1964.
The RTG burned up on re-entry with the rest of the spacecraft, as it had been designed to do. That increased the global environmental burden of plutonium and plutonium products by 4%, but the mission designers preferred a tiny increase in global risk to the large increase in local risk that would have occurred if a damaged RTG landed in a populated area. After that, the US reconsidered policy on the construction of RTGs, and designed them to survive re-entry and impact.
In 1965, the US launched their first and only space nuclear reactor, the prototype SNAP-10A. Reactors generate heat by means of a controlled fission chain reaction, instead of simple radioactive decay as in an RTG. The SNAP-10A operated for 43 days; at last notice, it was still in orbit. Later that same year, the USSR sent its first RTG-powered satellite into space.
There was another US RTG accident in 1968, when the "Nimbus B1" weather satellite took an unexpectedly short trip into the Pacific Ocean after a launch attempt from Vandenberg AFB. The satellite fell into shallow water off the California coast, and the valuable RTG was recovered intact, to be used on a later mission.
The third and last US RTG accident took place in 1970, as a consequence of the near-disastrous Apollo 13 mission. The Apollo Lunar Module (LM), used by Apollo crews to shuttle themselves to and from the Moon, carried an RTG as part of a lunar science package that the astronauts would set up and leave in operation. However, the Apollo 13 crew used the LM as a "lifeboat" to help get them back to Earth and the science package remained on board. The spacecraft was guided to a reentry over the Tonga Trench in the Pacific, one of the deepest oceanic canyons in the world, and no radioactivity was released before impact.
The United States gradually lost enthusiasm for nuclear power in space, due to environmental issues and the cost of building an RTG that could meet stringent safety requirements. RTGs were only used on US deep-space probes, such as the Viking Mars landers, the Pioneer 10 and 11 probes, the Voyager 1 and 2 probes, and the Galileo and Cassini probes. In the deep reaches of space beyond Mars, solar power wasn't generally seen as strong enough to power a sophisticated spacecraft and an RTG was the only practical option.
There was still public opposition to RTGs. In 1989, just before the launch of the Galileo Jupiter orbiter probes, anti-nuclear activists began protests against the launch of the spacecraft. NASA was by no means unprepared for such an attack, and in fact the agency had put a massive amount of effort into ensuring that the RTGs were as safe as possible, not to mention in preparing answers to almost every possible objection in advance. Mainstream environmental organizations regarded the matter as of such marginal significance that they did not join in to block the launch. The activist group, which included the prominent "monkey-wrencher" Jeremy Rifkin, took NASA to court, but NASA's won the case easily.
* The Soviet Union remained much more interested in nuclear power in space. Although the Lunokhod Moon rovers were powered by RTGs, the main rationale for Soviet space nuclear power was the military "radar ocean reconnaissance satellite (RORSAT)". RORSATs provided a round-the-clock, all-weather system to track western naval forces, providing targeting data to Soviet ships, submarines, and aircraft with long-range cruise missiles. Since the radar system on the RORSATs had limited range, these satellites operated at a relatively low orbital altitude of about 250 kilometers. At that height, solar panels would have dragged them down, so they were powered by small nuclear reactors that generated about two kilowatts of electricity.
The RORSATs were designed with a safety feature that in principle ensured they didn't dump radioactive materials back down on the Earth. When a RORSAT was withdrawn from service, a booster rocket kicked its reactor section to a high orbit where it would remain for hundreds of years while its radioactive material decayed to safe levels. The main body of the satellite then de-orbited and burned up.
Unfortunately, the RORSAT Cosmos 954 suffered a malfunction and could not kick off its reactor section. The satellite fell out of the sky on 24 January 1978 and scattered radioactive debris over large areas of northwest Canada, fortunately a largely unpopulated region. US President Jimmy Carter proposed a ban on atomic-powered Earth satellites -- but the Soviets, who had a real need for such spacecraft while the US did not, didn't respond. The embarrassment of Cosmos 954 led to a redesign of RORSATs. They now included a backup fuel-core ejection system to allow the core to burn up in the atmosphere. That would raise the global atmospheric radioactive load by a small amount but prevented any one location from suffering a rain of radioactive debris. In 1983, the RORSAT Cosmos 1402 underwent an uncontrolled re-entry, with the fuel core ejected and at least partially burned up over the South Atlantic Ocean.
There was a close call in April 1988, when the RORSAT Cosmos 1900 stopped responding to commands. Days before the satellite was due to fall to Earth, a backup system engaged and kicked the reactor to a safe orbit. Soviet Premier Mikhail Gorbachov then ordered the RORSAT program canceled. However, the year before that, the USSR had tested the next-generation "Topaz" space nuclear reactor, flying it on Cosmos 1818 and Cosmos 1867. After the fall of the Soviet Union in the early 1990s, the US expressed interest in the technology and bought an unfueled Topaz reactor for inspection.
* In fact, the US was very interested in orbiting large space nuclear reactors during the heyday in the 1980s of the Strategic Defense Initiative missile-defense program, as a means of powering large space defense stations, and worked intermittently on a 100 kilowatt orbiting reactor with the designation "Space Power 100 (SP-100)". It went nowhere; when SDI's tide ebbed, the need for SP-100 largely evaporated.
Public environmental concerns over space nuclear power made the whole idea a hard sell. The uncertainty was aggravated by resistance from the space astronomy community, since space atomic power systems in Earth orbit can confound high-energy astronomy satellites through emissions of gamma rays. The SP-100 program was shelved and space atomic power went on the back burner for about a decade.
By the early 21st century, NASA was becoming interested in atomic power again. Partly the revival in interest was because the agency was running out of plutonium-238 (Pu238) to fuel new RTGs; the US had stopped synthesizing Pu238 after the end of the Cold War, but was able to keep on obtaining from Russia. In 2009, the Russians stopped deliveries, apparently because they'd run out of Pu238 themselves.
That left inadequate fuel to power RTGs for projected deep-space missions. NASA, faced with a crunch, began work on radioisotope generator systems based on Stirling-cycle heat engines instead of thermoelectric junctions. A Stirling-cycle engine operates simply by making it hot at one end and cool at the other, with a heated gas driving a set of pistons. Stirling-cycle engines are more complicated than thermoelectric junctions, but they are about four times more efficient, which translates to a much cheaper and smaller RTG using a smaller load of radioactive materials.
That effort proved inconclusive. In 2013 the US Congress authorized the DOE to start synthesizing more Pu238; the supply bottleneck ended, NASA shelved work on Stirling-cycle converters. They were sexy, but NASA simply couldn't afford the luxury of bringing them up to operational service. We may not have seen the last of the idea, of course.
BACK_TO_TOP