How does electric space propulsion work?

Electric space propulsion moves a spacecraft by pushing gas out of the back at very high speed, using electricity rather than burning fuel. A chemical rocket gives a strong push for a short time. An electric system gives a gentle push for a long time, and that difference decides where it is useful.

It matters because a faster exhaust needs less propellant for the same change in speed. In 2018 the European Space Agency (ESA) said the exhaust velocity of BepiColombo's thrusters is 15 times greater than conventional chemical rocket thrusters. Less propellant can mean a lighter spacecraft or a longer mission. The examples below are deep-space probes that have flown.

How it works

The spacecraft carries a store of propellant. In the BepiColombo example, the propellant is xenon, an inert gas. Inside the thruster, inert xenon gas is fed in, and electrons are first stripped off the xenon atoms. The resulting electrically charged atoms, called ions, are then ejected.

On BepiColombo, a high-voltage grid system focuses the ions and ejects them out of the thrusters. Each ion is light, so the force from any one of them is small. Running the thruster for a long time adds up that small force into a useful change in speed.

BepiColombo and Dawn get their electricity from solar arrays. ESA describes twin solar arrays, each 15 m long, on BepiColombo's Mercury Transfer Module. The German Aerospace Center (DLR) says Dawn had solar panels with a span of 20 m.

What has been demonstrated

These results come from missions that flew and from agency records.

Deep Space 1 launched from Cape Canaveral on 24 October 1998 and tested 12 advanced, high-risk technologies in space. It was retired on 18 December 2001. Its 2.3 kW xenon ion engine provided primary propulsion and two-axis attitude control, thrusted for more than 16,000 hours and consumed more than 70 kg of xenon.

Dawn used three ion engines, and its ion engine used a maximum of 2,500 W. It entered orbit around Vesta on 15 July 2011 and Ceres on 6 March 2015, and its mission ended on 1 November 2018. At Vesta, it used its ion propulsion to spiral down to an altitude of about 210 km (130 miles).

BepiColombo, launched on 20 October 2018, relied on its electric propulsion system to reach Mercury. As of 23 October 2026, ESA's mission page lists it as in operation and en route to Mercury, with six Mercury flybys between October 2021 and January 2025 and arrival at Mercury due in December 2026.

What is only reported

These claims come from mission plans, project pages or news reports. This article has not verified them independently.

In November 2018, ESA said that BepiColombo's two power processing units were designed to fire two of its four T6 thrusters at once, guaranteeing a maximum thrust of 250 mN, and that the ions leave at 50,000 m/s.

A May 2019 news report from Spaceflight Now said that Starlink satellites carry Hall thrusters that use krypton, and that most satellites with Hall thrusters use xenon. The same report said krypton is less expensive than xenon but offers less thrust efficiency, citing a 2011 paper by US Air Force and satellite industry engineers. Satellite designs may have changed since 2019.

ESA's ARTES AT Kr project, listed as ongoing with a status date of 15 January 2025, aims to mature the cathode of Safran's PPS X00 low-power thruster for operation with krypton. The project page says all the elements needed to claim TRL6 maturity for that cathode are available. That is the project's own statement, and the page gives no numerical krypton efficiency.

What is speculative

The sources reviewed do not support firm predictions. ESA's ARTES AT Kr page says krypton operation is of particular interest for very large satellite constellations, where xenon availability is a key issue. Whether krypton cathodes from that project reach flight use is not shown in the sources.

Limits and open problems

Thrust is very small. ESA gives a maximum of 250 mN for BepiColombo's propulsion system, and says the thrusters have to keep firing for long periods to be effective.

Power is a constraint. The spacecraft needs large solar arrays, with two 15 m arrays on BepiColombo's Mercury Transfer Module.

Propellant is a constraint too. ESA's ARTES AT Kr project page names xenon availability as a key issue. Krypton is cheaper than xenon, but, according to the 2011 paper cited by Spaceflight Now, gives less thrust efficiency.

Lifetime is shown only for a few missions. Deep Space 1's engine ran for more than 16,000 hours, but this article found no verified reliability figures for the wider market. It also found no verified cost or mass figures across the market.

What to watch next

Watch the ARTES AT Kr project for results on krypton-operated cathodes, and for any verified efficiency figure for krypton on the same thruster. The project was listed as ongoing at its January 2025 status date, so its results may not yet be published.

Watch BepiColombo's arrival at Mercury, which ESA's mission page puts in December 2026. ESA's 2018 plan expected arrival about seven years after launch; the mission page now describes an eight-year cruise.

Watch for flight data from the satellite constellations that use Hall thrusters, since most public figures for those systems come from news reports and project pages rather than from measured results.