Purchasing new hardware? Read our latest product comparisons

Improved ion engines will open up the outer Solar System


March 6, 2013

An ion engine test for Deep Space One (Photo: NASA/JPL)

An ion engine test for Deep Space One (Photo: NASA/JPL)

Image Gallery (4 images)

The phrase "engage the ion drive" still has the ring of a line from Star Wars, but these engines have been used in space missions for more than four decades and remain the subject of ongoing research. Ion engines have incredible fuel efficiency, but their low thrust requires very long operating times ... and therein lies the rub. To date, erosion within such an engine seriously limits its operational lifetime. Now a group of researchers at NASA's Jet Propulsion Laboratory (JPL) has developed a new design that largely eliminates this erosion, opening the gates for higher thrust and more efficient drives for manned and unmanned missions to the reaches of the Solar System.

Ion engines of various types have been used on space missions since at least 1964, when NASA flew the suborbital Space Electric Rocket Test I mission. Many classes of space missions can benefit through using fuel efficient ion engines during some phase of their mission. For example, several communication satellites have been raised into their final geosynchronous orbit using ion thrusters. The European Space Agency's SMART-1 lunar mission was placed in geosynchronous orbit by conventional means, and then made the transfer into lunar orbit using an ion engine.

Deep space missions, however, is where ion engines could really shine. Three missions, NASA's Deep Space One and Dawn, and the Japan Aerospace Exploration Agency's Hayabusa have partially or entirely obtained their post-Earth-orbit propulsion from ion engines. Their ion engines operated for several years with only an occasional panic attack while providing a few hundredths of a Newton (perhaps 0.4 oz) of thrust.

How does an ion engine work?

There are many varieties and more proposals (the VASMIR engine comes to mind), but the operating principle is quite simple. There are two basic styles of ion engines, electrostatic and electromagnetic.

An electrostatic ion engine works by ionizing a fuel (often xenon or argon gas) by knocking off an electron to make a positive ion. The positive ions then diffuse into a region between two charged grids that contain an electrostatic field. This accelerates the positive ions out of the engine and away from the spacecraft, thereby generating thrust. Finally, an neutralizer sprays electrons into the exhaust plume at a rate that keeps the spacecraft electrically neutral.

An electromagnetic ion engine also works by ionizing a fuel. In this case a plasma is created that carries current between the ionizing anode and a cathode. The current in turn generates a magnetic field at right angles to the electric field, and thereby accelerates the positive ions out of the engine via the Lorentz force – basically the same effect on which railguns are based. Again a neutralizer keeps the spacecraft electrically neutral.

Powering a serious spacecraft

This all takes a good amount of electrical power – about 25 kW per Newton (3.6 oz) of thrust. So what thrust levels are needed to push, say, a 100 ton spacecraft throughout the solar system? (Forgive me – I like to dream!) It depends on the mission, of course, but 1000 N of thrust would put that spacecraft in orbit around Jupiter in about 10 months and in Neptunian orbit in just under 1.5 years. Obviously this is pretty far down the pike technologically, but let's see what is needed.

First, a supply of electrical power that delivers about 25 MWe (megawatts of electric power) more or less constantly. Clearly, we're talking about nuclear power – a lot of nuclear power from a reactor system that fits within a 100 ton spacecraft. Fortunately, there is currently a good deal of effort going into designing compact nuclear reactors for power production here on Earth.

In addition, NASA and DOE are collaborating on the Fission Surface Power Project, where the goal is to make tiny nuclear power reactors for bases on the Moon and Mars. The design goal is to produce a reactor that will provide 40 kWe for 10 years, fold into a 3 x 3 x 7 meter (10 x 10 x 23 feet) space and weigh in at 11,000 lbs (5000 kg). This is quite a ways from what is needed for a 1000 N ion drive, but with molten salt reactors and efficient conversion of heat into electricity, it seems within the bounds of possibility. Besides, I said I was dreaming.

Design roadblock

If the above were worked out satisfactorily, could we build a 1000 N thrust ion engine? There are some minor technical problems with efficiently ionizing the fuel and cooling the engines, but the biggest roadblock of which we are currently aware is that the large ionic current passing through the engine will cause enough erosion to destroy the engine. This is not a materials problem – it is a design issue. This is the roadblock recently demolished (at least partly) by NASA researchers at the Jet Propulsion Lab in Pasadena, California.

You can see in the cross-section diagram above that the fuel plasma fills the anode and gas distributor. At low thrust, the small plasma density is accelerated by the Lorentz effect of the crossed magnetic and electric fields. However, at large thrust, the plasma density becomes large enough to distort the fields, resulting in positive ions being accelerated directly into the anode walls.

When these ion energies are large enough, they will erode material from the walls in a process called sputtering. To make matters worse, in the quest for better ion engines, one desires both larger thrust and larger exhaust velocities (which require less fuel). Changes made to meet both of those goals greatly increase the rate of erosion.

This problem is made more difficult because the electrodynamics of the fields and the plasma are seriously nonlinear, making it difficult to predict the effect of a change in engine design on the erosion of the engine.

The obvious approach was to magnetically shield the walls from the energetic ions. The NASA team accomplished this by shielding the boron nitride walls so that the magnetic field from the inner and outer magnetic coil would pass around the end of the anode annulus. Properly done, the magnetic field no longer penetrated the walls. As a result, the magnetic field lines, rather than penetrating the walls at angles close to perpendicular, are nearly parallel to the walls. This causes the positive ions to be accelerated away from the walls, and as a result the walls are effectively the coolest part of the internal engine surfaces.

The result of experimental tests of the new magnetically shielded configuration showed the rate of erosion was reduced by a factor of 500-1000. This highly successful demonstration took place in a six kW Hall effect ion thruster.

While there will no doubt be more challenges to overcome as work proceeds in the further development of large-scale ion drives, this new research looks to have solved the nearest and most visible problem. Manned deep space missions are one step closer, and having dreamed of space travel all my life, I'd like to live to see it happen.

Sources: Applied Physics Letters, NASA-JPL

About the Author
Brian Dodson From an early age Brian wanted to become a scientist. He did, earning a Ph.D. in physics and embarking on an R&D career which has recently broken the 40th anniversary. What he didn't expect was that along the way he would become a patent agent, a rocket scientist, a gourmet cook, a biotech entrepreneur, an opera tenor and a science writer. All articles by Brian Dodson

It looks like that engine is just one small cylinder magnet and a wide thin cylinder magnet while the hole thing is run off of a Tesla coil for the output for the anode and ground for the cathode.

Travis Moore

While this is very encouraging, the real problem with space travel is the cost of getting materials out of the Earths gravity well. While this new technology will help with unmanned craft, I do not see a real long range manned system happening until we can launch every heavy sections of a craft to be assembled in space. I have a running bet with a Physics Professorial friend, I am betting on electromagnetic launch he thinks the space elevator is going to win the day. I have 1$ riding on it:)


The cost to orbit using chemical rockets is not as bad as you might think. The majority of cost is the cost of the rocket. If we can create a reusable rocket the fuel cost is quite reasonable. The fuel for a resupply launch to the space station only costs about $200,000. The big cost is in the rocket we have to throw away now. We are not talking about a Space Shuttle approach that was a total financial disaster costing more to refurbish after each launch than simple throw away rockets cost. We need a system that can use a simple low cost ablative heat shield for reentry and deacceleration of the booster until it reaches terminal velosity with a controlled fall through the air. With out all the fuel the crafts density is lower than that of a human being so terminal velosity is about 100 to 200 mph. When you get close to the ground restart the engine and make a controlled landing. Preferably near the next launch location. Refuel add a new heat shield and payload ready to go again. SpaceX is working on this and I believe they will succeed.


The author is very optimistic about space nuclear power's funding level, as of late the Fission Surface Power (FSP) Project has been zero funded following Obama's cancellation of the Constellation Project (from which it received its funding).

All of the nuclear systems he speaks of are PAPER reactors (meaning academic studies and computational work)! Very little has been done in terms of actually building and operating one of these little gems, which is unfortunate.

It's nice to talk about the engines that get you to the outer planets, but until the US concertedly funds a space nuclear project to power these devices, they are just an engine for making pretty lights on earth.

Gwyn Rosaire

MontyPython... How prevalent is elemental hydrogen in deep space? I'm assuming the electro-negativity of the ions is proportional to the sputtering problems (like nickel/platinum nucleate sites for brown's gas production). With a source of fuel in space it would only be necessary to magnetically(electrostatically?) push off of it for thrust (like a jet engine intake/exhaust). (recombining to regain fuel and repeat? hydrogen 1.2eV.) Passivating the material only delays the oxidation(sputtering). Maybe something like molygraphene(?molybdenum-?sulfide??) used for the magnets in the one shown in the pic. Might be lighter weight too.


Is an Ion engine the most efficient way to turn the energy of a nuclear reactor into thrust? The reactor generates heat. Then some mechanism (therocouples or a fluid driven turbine) converts that heat into electricity with efficiency losses. The ion engine + fuel then converts the electricity to thrust. Isn't there some way to combine the heat of the reactor with the ion engine fuel to get thrust and skip the heat to electricity and electricity to thrust steps?

Les LaZar

n2liberty, I love Space X and hope they do make their rockets reusable and self landing (they do more in a year than NASA does in a decade). I have been watching them closely and love what I see. However, rocket launching is dangerous and even if they land their rockets for reuse, you then have the issue of how many times do you reuse a part before you replace it or rebuild it. This leads to all new maintenance and safety issues. With electromagnetic launch you would still use a rockets. Very safe, low thrust rockets to keep accelerating the "truck" space vehicle into orbit. Once the vehicle drops off its load, it returns under powered flight. When I propose the idea people laugh and say "That would have to be huge!". Yes I am talking a couple miles of flat and then up the side of a mountain. We are talking billions (still cheaper than the Shuttle....) There would be only one. And the first country that made one would put everyone else out of business.....


Les, I completely understand where you are coming from. The problem is that each module Is designed separately then the modules are assembled together to form a working system. The efficiencies you are suggesting would imply that each project is designed completely customised using no 'off the shelf' modules. This would me insanely expensive. The modular approach is cheaper, but the "one module fits all" approach is never optimal.

Simon Sammut

Gwyn Rosaire, one of the nuclear reactor types that you say are are just 'Paper Reactors' is in fact not a 'Paper Reactor.' The Molten Salt Reactor Experiment at ORNL ran for several years and was poised for commercial viability until politics gutted the project. No paper designs and math here, it's a real, usable piece of technology.

Daniel Moreno

"When these ion energies are large enough, they will erode material from the walls in a process called sputtering. " Is the rate of sputtering proportional to thrust output?

Just an idea. Reaction chamber damage is not easily predictable. So why waste material reinforcing the chamber when you can just repair on the fly.

Can the sputtering effect be offset by a material that repaired itself. Or more simplistically, can a directional nozzle inside the reaction chamber spray a coating of material to repair walls affected by sputtering.

In this way you can run the ion drive at maximum thrust, and every few week/months/years power down, assess the damage, repair, and power back up again. Likely at a certain thrust level, there would be a distribution of errosion in just a few spots.


@Les LaZar: Sure. You can cut out the middle man entirely and just use the reactor core as a primary heat source. Store the reactor core inside a pressure vessel, and connect to it an exhaust jet. Just flush the toilet and off you go.

The problem however, is that we're now talking about a steam powered space ship. While that's a ridiculously cool idea (and I hope someone does it some day), getting around in space is all about weight to thrust ratio, and weight to thrust comes down to exhaust velocity and mass.

When discussing efficiency in propulsion, there's more at play than simply mechanical efficiency. We have to consider that not only do we need energy, but also fuel mass. Fuel mass ultimately is the most important consideration to propulsion in a vacuum, as is by far the most scarce, and the more you bring, the more you have to burn to move it. A nuclear/steam powered rocket nozzle might (i'm pulling numbers out my ass here) provide an exhaust velocity of 1km/s. I doubt it, but let's call it 10km/s just to be safe. Conversely, the theoretical limit to the exhaust velocity of an ion thruster, is limited only first by the energy available and eventually by the speed of light. 10km/s to 300,000km/s. Assuming even 10,000km/s, you require 1,000x less fuel than our totally unrealistic steam thruster. But that's not even true because you need fuel to push that 1,000x. I believe this works out to some magical math that has to do with inverse squares and some X's and Y's or something.

So in terms of mechanical efficiency, the nuclear steam engine is ridiculously efficient when compared to a nuclear generator providing electricity. In terms though, of propulsion efficiency, with fuel mass brought into the equation, the steam engine feels like a horse drawn carriage being compared to ... well, an ion thruster I guess.

Ben Selinger

PrometheusGoneWild.com - this is just a response to your earlier comment. Whilst they are both still at a very early stage, the asteroid mining ventures could eradicate the need to transport materials from Earth for space craft manufacturing. With their much smaller size, if asteroids and other space debris could be captured together with their kinetic energy, the resulting raw material facilities could end up completely self sufficient.

Senake Atureliya

re; Les LaZar

In general I agree with Ben. I also think nuclear rockets such as NERVA will be used for surface to orbit rockets for planets and moon other than earth with atmosphere.

Nuclear "Steam" rockets are fun to think about they are significantly less efficient than an ion rocket. Nuclear rockets run cold compared to other rockets their efficiency comes from the of free hydrogen as reaction mass. In a liquid hydrogen fueled chemical fueled rocket the maximum efficiency is achieved by putting a higher percentage of hydrogen in the mix than what will be burned leaving the extra hydrogen to leave the nozzle as free reaction mass. A nuclear reactor powering ion engines is more efficient. However the level is much higher so having the reactor core capable of performing as an atomic rocket for emergency evasion maneuvers is a good idea.


Is there a point to taking a nuclear reactor with you in a solar system supplied with a 864938km diameter fusion reactor ?. Build very big reflectors on your ship, concentrate sunlight at a gas turbine electrical generator, plug in the ION drive. Not much good for Pluto, but what is there there for a space tourist anyway, but by the time you reach Mars you have already got enough velocity to reach the outer solar system.


re; L1ma

You need power to keep your ship running in transit and cutting thrust at Mars orbit will add significantly to the length of the journey to the outer planets and you will need power for ships functions. Solar will not do.


Re Slowburn;

Take the indirect route; accelerate towards the sun, use a slingshot orbit while accelerating with the ION drive. Most people would consider this normal, Mars orbit is not the location of Mars but its orbital path around the Sun, at this distance 227,900,000 klm there is 1/3rd the available sunlight than at Earth Orbit at Jupiter 1/27th and at Saturn 1/100th. Consider that there is no limit to the size of the refector you can create in space, it is perfectley feasible to make one 100 times larger than necessary to power your generator at Earth orbit and power your spacecraft all the way to Saturn. Mylar refective foil weighs next to nothing compared to Plutonium.


re; L1ma

We currently can't get 3m x 3m Mylar reflectors to deploy correctly. How long will it take to learn to get your multiple kilometer x multiple kilometer reflector to deploy correctly.

Nuclear reactors work, are capable of providing for rapid velocity changes for emergency evasive maneuvers, and provide power when your ship is in planetary shadows.

How much does the many square kilometers of aluminized Mylar, the frame to hold it, and the batteries to power your ship when in planetary shadows weigh?


I forgot to mention the same radiation shield that will protect you from solar storms will shield you from your nuclear reactor.


Re slowburn;

Currently there are no reactors of the type specified in the article in service.

To get 40 kwe at earth orbit, sunlight strength at the Moon is currently 343 WattsM2 so we need 116 meters square of Mylar, at Saturn to get the same power we need 11600 meters square of Mylar, Mylar is listed at 6.8 gramsM2 so for a Saturn trip we would need 680 grammes (6.8 Kg) of Mylar, The generator used would be the same one for your reactor along with its radiators. The reflector would be stiffened with carbon carbon rods and operated with actuators(Parasol shape). I am trying to find a weight problem and if there is one it is that I come in around 4800 grammes lighter than your nuclear reactor(Not including generator). I have weight to spare for Lion batteries.


re; L1ma

Ignoring that we can not currently get a 116 meter square aluminized Mylar mirror to focus the light in the right place, How many passes through Saturn's rings will your aluminized Mylar mirror survive?

Also instead of a tiny little probe, think of a manned spaceship.


Correction 4800kg's lighter


Re Slowburn.

"How much does the many square kilometers of aluminized Mylar, the frame to hold it, and the batteries to power your ship when in planetary shadows weigh?" "Also instead of a tiny little probe, think of a manned spaceship." You got your answer which is equivalent for the study in the article for a 100 tonne spacecraft.

"Ignoring that we can not currently get a 116 meter square aluminized Mylar mirror to focus the light in the right place" There is no requirement for prefect parabolic focusing with a mylar mirror, never will be - read up on solar furnaces. "How many passes through Saturn's rings will your aluminized Mylar mirror survive?" Unfurled Mylar will not survive passing through the rings of Saturn which consist of everything from snowballs, ice to meteors. I never said they would, the horse will get you to your destination. By the same definition I doubt your nuclear reactors radiators would survive either, they need to be exposed, being holed once would mean total failure with coolent loss.


re; L1ma

The final destination is to return to earth.

The orientation of the radiators is unimportant giving plenty of time to aline for minimum interference where as the aluminized Mylar must be keep alined for solar collection. I can take a nuclear powered craft through the divides or under the rings neither of which is possible with your aluminized Mylar reflector. In fact you could park a nuclear powered ship in the ring without taking too much damage because everything hitting you will be at a similar velocity.

The coolant in the radiators would be a extremely low vapor pressure liquid in fact using molten lithium the liquid coolant can be the radiating surface without excessive loss of coolant so a small puncture will not result in mass coolant loss.


Re Slowburn, The nuclear powered craft currently suitable for entry into a debris strewn environment would be those powered by radioisotope decay, not active fission reactors. Lithium salts would not be suitable, though transfering heat well as a liquid and part of the reactor design, IF the radiator is holed, the salts would leak out and solidify around the craft, and lithium cannot power a turbine, the flow rate is too low. The coolant Li2Bef4 is used in liquid fluoride reactors and reaches temperatures of 700 degrees and requires a minimum of 345 degrees to be molten, flow rate is usually at 120 cubic meters per second and uses a heat exchanger to power a steam turbine, whose exhaust is condensed to liquid via another heat exchange - in your case the reactor radiators. In space I would expect helium gas instead of steam. This still leaves the radiators as damage prone as mylar. There was some work on cooling with lasers, making a 40 degree drop in temperature possible (un Zhang, Dehui Li, Renjie Chen, and Qihua Xiong) using cadmium sulphide which if you used the current solid state laser technology would make your reactor safe to use for it would require no exposed elements.


re; L1ma

Why would I need to power a turbine? Stirling cycle is more efficient. I said lithium not lithium salts. while I might use lithium salts in a heat-sink I will use pure lithium with its melting point of 180.54 °C as the coolant. My radiators are vastly smaller than your Aluminized Mylar reflector and made out of materials vastly more durable.


Re; Slowburn,

Stirling Engines are listed as under development under hardware risk reduction and are not slated in as the sole option which at this time is a Brayton engine. It may be a combined cycle with Brayton and Stirling. Good luck with your Lithium salts.


In case you missed it, pure Lithium is a highly reactive alkali metal which will react at its transition temperature with the reactor components to form amalgum, unfortunatlely current Stirling engines and the Brayton engine only work with gases, not liquid metals and cannot directly generate power from them . Lithium needs to be safely combined with floride to become inert chemically I think you are confusing Thermocouples which would do this job directly without having to use heat exchangers(Soviets used RTG's frequently)


Adding to this, the mission profile would be very different to a solely nuclear craft, the possibility is there to have up to 5 or more energy collectors to power your ION drive for the same mass at mission start, giving an intial 5000Newtons of thrust if you take several engines - ION dives have minimal mass. Mission time to the destination - say jupiter would come in at 1/3 rd of the stated 10 months, the craft would have to hover under power above the plane of the ecliptic in Jupiter orbit, say above its pole and an RTG powered orion would then have to drop into the Jupiter system under its own power.


re; L1ma

Metallic lithium. Lithium salts are best for storing heat and much more difficult to pump.


re; L1ma

It must be possible to contain molten lithium given that this document exists. http://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/19680018893_1968018893.pdf

Thermocouples are particularly susceptible to neutron damage, heavy for the electricity they produce, and must have a thermal difference to generate electricity so you will still need a radiator.

The reactor I would use is an updated NERVA engine and while I would xenon as the primary core coolant but I would also a carry hydrazine as reaction mass for landings/launch on moons and emergency evasion maneuvers. This will require a significant reserve of xenon but less than than the hydrogen that would be lost using hydrogen when using the reactor solely for electricity production.

Why would you hover on thrust rather than orbit Jupiter? On Saturn it might be defensible.


The whole thing is theoretical BS. Until such time as there is a solution to the problem of inertia, humans will never be leaving the solarsystem.

A human can not live in an environment with an artificial gravity greater than or less than Earth. That means any extended voyage has to accellerate at 9.8 m/s/s. Giving you an artificial gravity equivalent to Earth. Unfortunately this results in whoever is on the space ship assuming you could build one large enough, would no doubt be extinct by the time it reached the next solar system let alone any other solar system and definately not another galaxy.

I would recommend you stop wasting your time and resources and focus on saving the planet because if you don't save this planet now you will not be around to build any technology to go anywhere in the future. Humans are not ready for inter-stellar travel and most probably never will be.


Now it will not do to try to concentrate sunlight far from the sun. Instead concentrate sunlight close to the sun , convert that energy to microwaves and tight beam microwave to your space ship, should have a much greater concentration of power per meter squared and the receiver does not need to be solid. Could also be used to power multiple things provided each had an attenna and they did not all need power at the same time.

Michael Hertel

This is very interesting. I would like to know the velocity and/or equivalent temperature of the exhaust. Is there any possibility at all of accelerating protons to a high enough velocity so the p+Boron fusion reaction is realized? Probably not but I was just wondering how far away this is from being a possibility? For the uninformed this is a nuclear reaction that produces no nasty radioactive byproducts. Dreaming on....


What's the range of thrust-to-weight ratios for the bare engine?

Andrew Palfreyman
Post a Comment

Login with your Gizmag account:

Related Articles
Looking for something? Search our articles