Ion Thrusters: Not Just For TIE Fighters Anymore

Spacecraft rocket engines come in a variety of forms and use a variety of fuels, but most rely on chemical reactions to blast propellants out of a nozzle, with the reaction force driving the spacecraft in the opposite direction. These rockets offer high thrust, but they are relatively fuel inefficient and thus, if you want a large change in velocity, you need to carry a lot of heavy fuel. Getting that fuel into orbit is costly, too!

Ion thrusters, in their various forms, offer an alternative solution – miniscule thrust, but high fuel efficiency. This tiny push won’t get you off the ground on Earth. However, when applied over a great deal of time in the vacuum of space, it can lead to a huge change in velocity, or delta V.

This manner of operation means that an ion thruster and a small mass of fuel can theoretically create a much larger delta-V than chemical rockets, perfect for long-range space missions to Mars and other applications, too. Let’s take a look at how ion thrusters work, and some of their interesting applications in the world of spacecraft!

It’s All About Specific Impulse

Chemical rocket engines provide huge thrust but are thirsty when it comes to fuel.
Ion engines won’t get you out of Earth’s gravity well, nor do they work in the atmosphere, but become useful when you’re in the vacuum of space. Credit: NASA, public domain

Before we dive into the world of ion thrusters, it’s important to understand the concept of specific impulse and fuel efficiency for rocket thrusters of all kinds. Specific impulse measures how effectively a rocket engine creates thrust from the mass it throws out the back, whether by chemical or any other means. The higher the specific impulse of a rocket thruster, the more thrust it generates per mass of fuel.

Impulse is the integral of force over time, measured in Newton-seconds. Specific impulse, where we look at impulse per weight of propellant, is thus measured in Newton-seconds divided by Newtons, or simply seconds. It’s a little confusing to wrap your head around, but for the newly initiated, just keep in mind that higher numbers of specific impulse stand for greater fuel efficiency.

For comparison’s sake, the Space Shuttle’s Solid Rocket Boosters get a specific impulse of just 250 seconds, while liquid oxygen-liquid hydrogen rocket engines may reach closer to 450 seconds. Electrostatic ion thrusters are almost an order of magnitude better, on the order of 2,000-3,000 seconds, with some reaching closer to 10,000 seconds in experiments, while the experimental VASIMR electromagnetic ion thruster predicts a specific impulse up to 12,000 seconds.

This better fuel efficiency has real implications for space travel. It means that an ion thruster can achieve a given change in velocity for a space craft with far less fuel – an order of magnitude less, in some regards. In an application regarding orbit-keeping for the ISS, one calculation suggested an ion thruster could reduce the space station’s annual fuel use from 7,500 kg to just 300 kg. This has flow on effects, where launch vehicles carrying that fuel to the space station need less fuel themselves to boost it into orbit, improving efficiency across the board.

How Thrust Via Electricity Works

Ion thrusters come in a variety of forms, but the basic principle is a simple one: electricity is used to accelerate ions to a high velocity, forcing them out of the thruster, thus resulting in a reaction force which propels the spacecraft itself. A neutral gas is used as fuel, which is ionized by stripping electrons from the atoms, resulting in a supply of positive ions that can readily be accelerated by electrostatic or electromagnetic means to generate thrust. Xenon, krypton, or argon are common choices for these thrusters, though other materials, like magnesium, zinc, and iodine have been experimented with in some designs. The vast majority of ion thrusters rely on gaseous propellants, however.

Electrostatic Thrusters

A schematic of an gridded electrostatic ion thruster. Wear on the grids over time limits the life of these thrusters. Credit: NASA

Electrostatic ion thrusters use a variety of methods to accelerate ions to generate thrust. Gridded electrostatic ion thrusters are one of the more popular designs, where the propellant gas is bombarded with electrons to form an ionized plasma. A set of gridded electrodes are then charged with a potential difference, accelerating the positive ions out of the thruster. A separate cathode then discharges low-energy electrons into the exhaust stream of the thruster to ensure the spacecraft doesn’t end up with a net negative charge.

Hall Effect thrusters replace the gridded electrodes with a gas-distributing anode and a magnetically-confined electron cloud acting as the cathode itself. The heavier positive ions are accelerated out of the thruster, while the more lightweight electrons remain confined in the magnetic field. Similarly, a external cathode is used to neutralize the exhaust stream as in the gridded thruster designs.

A schematic of a Hall Effect thruster. Hundreds of such thrusters were used for stationkeeping in Soviet satellites in the 20th century. Credit: Finlay McWalter, public domain

These designs have seen significant use in real-world missions. One of the earliest applications was in Soviet satellites, which used Hall Effect thrusters instead of chemical rockets for station keeping. This is where satellites need to periodically apply thrust over time to counteract the subtle atmospheric drag they experience. The miniscule thrust provided by the Hall Effect thrusters is fine for this purpose, applied over a long period for a significant overall change in velocity.  Power draw of these thrusters was on the order of 1.35 kW, generating 83 mN of thrust for a specific impulse of around 1,500-3,000 seconds.

A more recent application of the technology is on the Chinese Tiangong space station, which uses four Hall Effect thrusters to maintain its orbit over time. NASA also hopes to fly the technology on the upcoming Psyche spacecraft, which will use four SPT-140 Hall Effect thrusters.  Loaded with 922 kg of xenon propellant, engineers have estimated that 15 times as much propellant would be required if Psyche relied on chemical rockets instead.

An SPT-140 Hall Effect thruster under testing. Four of these thrusters will be installed on NASA’s Psyche spacecraft. Credit: NASA, public domain

Gridded ion thrusters have seen plenty of use, too. NASA’s NSTAR ion engine was installed on the Deep Space 1 probe, which was sent out to fly by a comet and asteroid in the late 1990s. The gridded ion engine put out just 92 mN of thrust for 2.1 kW of power, but its high specific impulse of 1,000-3,000 seconds enabled significant mass savings compared to a chemical rocket solution for its interplanetary journey.  The ion thruster, fueled by xenon gas, ran for a total of 16,265 hours during the mission, providing a total change in velocity (delta-V) of 4.3 kilometers per second, the largest for any spacecraft relying on its own onboard propulsion system.

Other deep space missions have also relied on the technology. JAXA’s Hayabusa probe relied on an ion thruster to help it rendezvous with the Itokawa asteroid. NASA’s Dawn mission also used the technology, being fitted with three of the same xenon ion thrusters used on the Deep Space 1 program, though only firing one at a time in practice. NASA was more than willing to point out the low thrust available from the propulsion system, noting that 0-60 mph would take four days, which compares poorly to the 3.5 seconds achieved by the average modern Ferrari.

Electromagnetic Thrusters

A prototype magnetoplasmadynamic (MPD) tested by NASA. Credit: NASA, public domain

Electromagnetic ion thrusters generate their thrust from neutral plasma, ostensibly consisting of equal numbers of positive ions and negative electrons, and are often referred to as “plasma thrusters” in literature. They come in a variety of designs, most of which use radio energy to ionize gas in a chamber. A magnetic field is then generated to accelerate the overall-neutral plasma out of the thruster. These designs often have the benefit that they don’t need special neutraliziation electrodes to correct the charge imbalance of the exhaust, nor do they use electrodes in the gas stream to accelerate the ions, reducing a source of wear in comparison to electrostatic designs.

One of the most well-developed examples is the VASIMR VX-200 thruster, which has been in development since 2008 in various forms by the Ad Astra Rocket Company. The aim is to operate the thruster at a power level of 100 kW for 100 hours, to indicate how the thruster can generate a huge delta-V for long-term missions. In July 2021, the company reached a milestone of 82.5 kW for 28 hours. The thruster performs with an exhaust velocity on the order of 50 km/s, with a specific impulse of around 5,000 seconds.

Electromagnetic designs often promise larger thrusts than electrostatic thrusters, though most are still in the research stage.  Issues with such designs include issues of high power draw and problems of dealing with waste heat. If these could be overcome, designs like a scaled-up VASIMIR electromagnetic thruster could propel a spacecraft from Earth to Mars in just 39 days, compared to the six month journey of a conventional chemical rocket. The only thing is, you’d need a power supply capable of delivering somewhere in the realm of 10 to 20 megawatts of power, and fit that in a spacecraft.

Looking To The Future

Ion thrusters in their various forms are in some ways a technology that haven’t yet proven their full capability. They’ve already done great things, taking small space probes to far-flung destinations while requiring far less fuel along the way. However, we’re still a long way from using them to help us get humans to destinations beyond our own orbit. There’s plenty of development still to happen before you’re riding an ion-powered craft on your future space holiday, but in 50 or a 100 years or so, an ion craft might just be the hot ticket to Mars!

39 thoughts on “Ion Thrusters: Not Just For TIE Fighters Anymore

    1. It’s awfully nice not to have to carry your propellant with you – this toy demonstration uses an electric field to push ambient air around.

      Hardly fair to call this an ‘ion propulsion’. Yes, it produces ions, which are driven by an electric field, but the vast majority of the force produced comes from the ions being dragged through air, imparting momentum to neutral air molecules.

      It’s just a horribly inefficient air blower.

        1. Leaving aside the implications of Newton’s laws regarding needing something to push against, it’s pretty impressive for an air ion engine to be able to lift its own weight directly. The best I’ve seen is a winged model airplane powered horizontally by a grid under its wings.

      1. from my calculation this vehicle is in the 3 to 5gramm per Watt thrust range.
        Most propeller drones in there normal operation area are in the 6-8 gramm per Watt range, so its not that horrible, still not in the 13-15 g per Watt range into which can be optained with a very efficient moter propeler combination but a good start i would say.
        And wait isn’t that the point of an ion thruster to impart momentum to the device by moving ions around? The point is all ion propulsers in the end eject a overall a neutral atom collection. Otherwise the the device would charge up and pull the charged atoms back, and no momentum from this mechanism would be generated.

  1. No mention of Starlink’s famous Hall-effect ion thruster? Nor how it failed to provide enough thrust to save a few dozen Starlinks from their premature re-entry deaths last month?

    I have not found a clear description of what, exactly, Starlink uses. But by the specs it’s very similar to Apollo/Ad Astra’s Ace Max: https://apollofusion.com/datasheets/Apollo_ACE_Max_Datasheet-Jan_2021.pdf

    That’s a 20 kg system mass, and can carry 10 kg of propellant, providing 50 mN of thrust for about 1000 hours. A total impulse of 180,000 Ns, and a total delta-V of 720 m/s on a quarter-ton Starlink.

    In operation, that thruster will eat basically all the electrical power available on the satellite.

    For comparison, using conventional hydrazine monopropellant instead (at 220s Isp), you’d need 83 kg of propellant, but much lighter system mass, especially when you don’t need to budget in additional solar panel size.

    I expect at this scale it’s a wash for which system is ‘better’. What is preferable from a handling perspective: a tank of toxic hydrazine you can walk away from if you drop it, or a COPV that can go boom and kill you promptly?

    1. SpaceX designed their satellites for low total cost, which includes launch cost. They were NOT designed to operate under the conditions of a solar storm occurring while they were at the altitude they were at, heating up and expanding the Earth’s atmosphere. And it wasn’t just the orbit; according to Scott Manley, the satellites were also reoriented to protect their electronics from the charged particle hail, and this put them in a higher drag configuration. This was a very low probability event. Not bad design, just very very bad timing.

      Also, hydrazine isn’t just toxic, and people don’t always walk away from accidents. https://en.wikipedia.org/wiki/1980_Damascus_Titan_missile_explosion (Also boom, when hypergolic fuels mix when they oughtn’t.)

      1. SpaceX’s press release and Scott Manley both say they reoriented the satellites to fly edge-on to minimize drag. There’s no mention they were “reoriented to protect their electronics from the charged particle hail,” and it’s unlikely they would have bothered: Orientation would not affect the dose rate very much, and a few hours of modestly increased dose rate during the storm would not count much against the years of expected lifetime. Atmospheric drag was the danger here, and they worked to minimize that.

    2. Those thrusters are DESIGNED not to provide that much thrust, because there’s an unbreakable tradeoff between specific thrust and absolute thrust. If you have to optimize for one, the other is unavailable due to physics.

      For practical satellite systems, there are a range of options:

      Full ion systems (very high mass efficiency, eat a lot of power, and take *weeks* worth of orbits thrusting in-plane during periapsis to raise apoapsis much (or the opposite if you’re trying to circularize afterward)

      Resistojets are lower specific impulse, pull less current, get less total delta vee out of the same fuel, but can be used for practical orbit changes when you can’t afford to wait 6 months for final orbit.

      Lots of other options across the spectrum.

      At the low end are monopropellant systems, those use a lot of fuel mass for much less total thrust (measured as the total change if you use all the fuel up). They do have *significantly* higher instantaneous thrust, although they’re embarrassed by pretty much every bypropellant system out there. They’re only favored because the plumbing and ignition are trivially simple and most monopropellants are “storable”: stay liquid, at moderate tank pressure across the large range of temperatures expected in space. A cryoliquid fuel ALWAYS beats them, but the tank will rupture or the fuel will boil away through the pressure vents after only a few months in orbit.

      The SpaceX design is mostly designed to counter *normal* atmospheric drag (at the higher final orbit, not at the really low orbit used for initial testing to ensure that uncontrollable satellites are rapidly removed by drag).

      This requires high specific impulse, and very small mass and volume dedicated to fuel. As such, it’s probably closer to a full ion thruster than a resistojet to maximize the mission time for a given quantity of fuel mass.

      At mission altitude, very little thrust would be needed, and that only intermittently, to keep the orbit correctly circularized and timed even in the face of solar-storm level atmospheric pressure variations (drag-at-height changes)

      But, while SpaceX took a gamble with the launch timing (solar event was already known to be likely), they have made the correct design tradeoffs: Initial orbital injection is kept very low to ensure that even a completely dead sat reenters in days-to-weeks, and once they check out as controllable and with working radios and power, their altitude is slowly raised to mission altitude.

      Remember, a failure at mission altitude will be cleaned up by drag in a couple of years at worst, but a failure at initial altitude will be cleaned up in days. It’s an anti-space-junk-in-useful-orbits strategy.

      If you want ALL the math and power/mass/thrust/time tradeoffs, start with the Wikipedia articles on the topic, then move on to the open-source professional literature on the subject.

      Space systems is such a broad field of physics and engineering that competence in any one segment easily creates false confidence in other areas. Just the practical differences within ion thruster families is extreme enough to cause this problem. For example, gridded thrusters tend to have effectively infinite practical wear lifetimes (if supplied with fuel and power), but hall-effect thrusters erode their chambers badly and fail after only a few thousand hours of runtime. It’s a field where nuances have huge impact on practical usage.

  2. the limit is actually space power systems. there are a lot of “works in the lab” ion thusters that could fly in space but are in want for a megawatt space reactor. you might say go solar. but the 250kw of power the iss has at its disposal (mind you the power system weighs in at over 9 tons). then you wont get much power past mars. even space nuclear is going to require a lot of radiators, and the reactors flown thus far kind of suck in terms of power output. ion drives are still more than adequate for unmanned probes, but its not the ass hauling man rated rocket you are looking for. move along.

  3. As a kid I remember my grandparents having some sort of encyclopedia or similar set which had an article about ion thrusters or ionic propulsion or something like that. I’m guessing this was from the 1950s. Anyway, I remember this article having a project in it. I may not be remembering it perfectly as I haven’t seen it since I was a kid. I remember it having a thruster built with a number of metal rods with metal balls on one end. I think they specified what metals should be used. This was hung by it’s power wires from a wooden stand. The whole thing was connected to a then-already-old automotive ignition coil that one would probably have to steal from Jay Leno’s garage today.

    When you connected power it was supposed to swing by ionizing the air around it so no Xenon supply was needed. Obviously it wasn’t going into space. I think it was presented as a science fair idea.

    Anyway, I remember really wanting to build it but not really being able to get the materials. Later I went looking for it and could never find it again. If this sounds familiar to anyone, if anyone knows what book set this was in that would be great. I want to say theirs was a World Book encyclopedia set but I guess it could have also been Brittanica. I think they had some sort of yearly supplement books too that were dated rather than lettered and may or may not have been the same company.

    I’m sure I could find something similar but it woudln’t be the same. As an adult with a job and access to the great big market that is the internet I like building those things I wanted to do but couldn’t as a kid.

    1. It was in the World Book Encyclopedia. Main edition, not one of the supplements. I remember reading it in the early 70s but I’m sure it goes back further than that.

      The one I remember had a push button to apply power. The idea was that the “rocket” was at the end of a pendulum. You’d push the button to give it a periodic impulse, kind of like pushing a swing. I suppose you could put it on a carousel if you could overcome frictional forces.

      I don’t remember the ignition coil or how much power it required but if all that’s needed is high voltage I wonder if a flyback transformer would work as well.

      1. I think at the time, Ford Model T ignition coils, which had an integral “buzz coil” to interrupt the primary current rapidly to generate a continuous spark, rather than using contact points, were popular for high voltage projects. They were easy to find, and all you needed was a 6V battery to power them. They were very distinctive, built into a wooden box. These days a flyback would be easier to find, and you’d have to drive it with an appropriate frequency.

        1. Yes it was Model T Coil.. I always wanted to build one. The T Coil is in a wood box.. There are four used in the Car.. One per Spark plug.. It has a Built in Buzz Circuit.. It initially was Driven by it’s ‘Horse Shoe Alternator’ that produced anywhere from ~4 Volts to 36 Volts AC depending on Engine Speed. Then a 6 Battery Circuit was added later to the T for Starter and Lights when they became available..
          The T Coil will run Nicely on anywhere from 6 to 24 Volts.. Just keep the Duty Cycle time down and Monitor the Case Temps.. Usually it’s the Internal Capacitor of the T Coil Box that can’t take the Cycle Times.

          1. The other issue (as an owner of a whole shelf of model t coils) is that the more or less integral capacitor fails through age, probably because it’s made of like waxed paper and copper foil or something, and then you have a pretty finger-jointed wood box full of potting compound that doesn’t do anything.

  4. Framing specific impulse in terms of weight, rather than mass, of fuel always seems weird to me – especially for things which will never be used near enough to earth for the reference gravity to have any meaning!

    Yes it gets you a nice unit of seconds, but does that actually correlate to anything useful? (possibly it’s proportional to how long something could hover for in standard gravity, as a multiplier of a function of fuel mass ratio?)

    Ns/kg instead reduces to units of velocity, and is in fact equal to effective exhaust velocity. So calling the mass-specific impulse mIsp, you get deltaV = mIsp * ln (wetMass / dryMass)
    So Ns/kg seems a lot more useful (for a fixed fuel fraction, there is a dimensionless linear scaling between Ns/kg and delta V), and more scientific (not using g in situations where g isn’t a relevant physical parameter)

    1. I completely agree with you, but all the armchair rocket scientists will come out of the woodwork here and cry “but that’s the way it’s always been done”, handwave a “conversion constant”, and complain that working in slug-seconds/poundmass is just too hard.

    2. I think it’s just an anachronism that never got fixed. In the early days of rocketry, efficiency of the (single-stage) engine was all that mattered, and that DID operate close to Earth. Remember that these guys used slide rules, so knocking a constant out of the calculation was a worthwhile thing. Now it’s just a customary unit. But hey! There’s no conversion required to go from pound-seconds/pound to kg-seconds/kg, so no chance to screw that up, either.

    3. The main advantage is that there’s a whole boatload of math that you’d have to rework, then recertify to NEVER be different under ANY unexpected corner condition…

      As an incidental, but insufficient for justification, it also makes a few of the very commonly used equations notably simpler (but others are more complicated, of course)

  5. Why does the spacecraft need to remain neutral? “A separate cathode then discharges low-energy electrons into the exhaust stream of the thruster to ensure the spacecraft doesn’t end up with a net negative charge.”

    Iunno if I’m just being foggy or what. Is the concern that the ship becomes negatively charged then the ions may be attracted to the ship or otherwise wouldn’t leave as fast as they’re attracted?

    1. Exactly: If the craft ejects only positive ions, it will quickly become the strongest negatively-charged object in the vicinity, and those ions are going to come right back home, giving back all that momentum and energy you threw them away with, negating your thrust. It just ends up being a complicated way to heat the exterior of the spacecraft.

  6. You could have mentioned the FEEP technology too, since there are commercial products available which are also successfully operating in orbit (currently around 100 thrusters AFAIK).
    They are a subtype of electrostatic thrusters but use liquid metals as propellants which makes them quite powerful and easy to handle.
    See:
    https://en.wikipedia.org/wiki/Field-emission_electric_propulsion
    https://www.enpulsion.com/ (a manufacturer of these thrusters)

  7. If I’ve read the wikipedia article on radiation pressure properly, 1 kW of electromagnetic radiation will produce about 7 micro-newtons of force. That’s about 4 orders of magnitude less per kW than an ion engine. However, the equivalent rest mass of a photon is extremely small, so the specific impulse should be quite high – about 3×10^7 seconds according to one source. Point LEDs out the back of your spacecraft and accelerate very slowly.

    1. 6.7 μN of force ONLY if the photons are reflected (like sunlight on a solar sail), and you get double the momentum. If you’re throwing brand new photons out your backside, you’ll get just P/c = 3.3μN.

      You don’t need LEDs. Just plain old radiant heat will work.

  8. I just wanted to mention, that the craft shown above has flown before carrying 2 small optional propellant tanks. There are videos of that on my Ethan Krauss YouTube Channel. The craft will therefore likely work in space also. With additional thin film solar cells, it should accelerate rapidly for quite a while. It just adds electrons to O2 molecules, so it can fly with less energy input.

    1. Cool. What’s the expected thrust/mass ratio of the complete assembly, with solar cells? Net delta-V? Oxygen is generally acknowledged to be awful for an ion drive propellant. How would the numbers change if you use a more suitable high molecular weight, low ionization energy propellant (Xenon, Iodine, liquid salts)?

      Akin’s Laws of Spacecraft Design #1: Engineering is done with numbers. Analysis without numbers is only an opinion.

      1. There are thin film Perovskite solar cells that produce about 30 watts of power per gram, so the craft would accelerate at nearly 1.4g. It requires exceedingly small amount of O2 at this point. Since it just adds electrons to the outer orbitals of the O2 molecules, it requires very little wattage. Xenon requires much more power because there is a specific amount of ionization energy required. I don’t know the exact propellant consumption because I’m a bit bogged down with the control system currently and want to finish that before doing vacuum tests. I do know it produces almost no lift in a pure N2 atmosphere and almost the same lift in a pure O2 atmosphere. That would imply only a very small percentage of the O2 gets ionized.

        1. Check your numbers.
          Ionization energy of oxygen is 13.6 eV = 1312 kJ/mole = 82 kJ/g.
          If you don’t hit O2 with too much energy it will remain a diatomic molecule, so you could see the diatomic ionization energy of 12.06 eV = 1164 kJ/mol = 36 kJ/g.
          Xenon’s ionization energy is 12.13 eV = 1170 kJ/mole = 8.9 kJ/g.

          For comparison, exhaust gas with a velocity of 3 km/s (Isp = 300 s) has a specific energy of 4500 kJ/g. In a well-designed ion thruster, ionization energy should be a small fraction of the total power budget. More important is how long your electrodes or other parts last. Atomic oxygen is not friendly.

          Now, figure out the actual power of your lifting force compared to your input power of the air lifter. Determine (or measure) the air velocity. What’s the specific impulse? What does the rocket equation say about your attainable delta-V?

  9. I dunno why they don’t have the ISS hovering on it’s own pee, so to speak.

    Now if you were going away from the sun I’d say you’d want humungous solar sails that were also a parabolic solar concentrator, and skip that weakass photoelectric solar panel efficiency trap by using most of the energy directly to heat a plasma, then have some of the most effective for PV photons prismed out for the panels to accelerate it electrically.

    The reflector should also be made of concentric panels such that it can be used to focus light from it’s rear for the return journey, reverse Fresnel style…. this would also “set the sails” to provide zero to negative photon pressure to the direction of travel.

    In return mode, it’s faintly possible it could also be directing free hydrogen for collection to augment propellant.

    1. Yeah,
      The ISS throws away, as trash and waste gas, about a kilogram of mass per hour.
      Coincidentally, it also burns about a kilogram per hour of fuel to maintain orbit.

      If you just throw out the trash at 3 km/s (Isp of 300), then you would not need to lift any fuel for stationkeeping.

      At 3 km/s the trash will promptly de-orbit, hitting the atmosphere at almost 5km/s: it’s still going to burn up, if the plasma/gasification process didn’t already disintegrate it.

      1 kg/hr at 3 km/s would only require about 2 kW of electric power: only a few percent of the ISS power budget. It’s crazy NOT to do dispose of trash as propellant.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.