1. The short version / what to know first / where this fits in the series

• Electric propulsion is a "small push, little propellant" method. It uses electrical energy to accelerate propellant and eject it. To change speed by 3 km/s, for example, a chemical system with a specific impulse of 300 s needs propellant worth about 64 % of the spacecraft's mass, whereas 1,500 s needs about 18 % and 3,000 s about 10 % Our calculation
• Thrust is set by power, and it is small. NASA's report says that electric propulsion generally gives a thrust-to-power ratio below 75 mN/kW, and that even a small spacecraft able to deliver 500 W produces no more than 38 mN. So the thrusters run for hundreds or thousands of hours Sourced[Ref. 1]
• The materials story comes down to "erosion, deposition and heat". The grids of an ion thruster, the channel walls of a Hall thruster and the impregnated material in a cathode must survive tens of thousands of hours of ion bombardment and high temperature. One life test ran to 51,184 hours and 918 kg of xenon Sourced[Ref. 8]

2. Why electric propulsion: specific impulse and the rocket equation

Throwing propellant backward pushes a spacecraft forward. The force is the thrust; NASA Glenn's explanation puts it as the mass flow rate of what is ejected multiplied by its speed Sourced[Ref. 2]. The same page defines specific impulse, Isp, as thrust divided by the weight flow rate of propellant; its unit is the second (s), and it is the same thing as the exhaust speed divided by the gravitational acceleration g0 Sourced[Ref. 2]. In other words, specific impulse is a fuel-economy figure: the higher it is, the less propellant is needed.

What does a high specific impulse buy? The propellant needed to change speed follows the ideal rocket equation Sourced[Ref. 3]. The propellant's share of the initial mass is 1 − exp(−Δv / (g0 × Isp)), where g0 is 9.80665 m/s² Sourced[Ref. 4].

As the figure shows, for Δv = 3 km/s the propellant falls from about 64 % (Isp 300 s) to about 18 % (1,500 s) and about 10 % (3,000 s). For Δv = 6 km/s it is about 87 %, 33 % and 18 % Our calculation. Chemical propulsion delivers a large force for a short time; in NASA's table the specific impulse of small-spacecraft chemical devices spans about 40 to 310 s. Electric devices span about 20 to 3,200 s depending on type Sourced[Ref. 1].

3. Thrust is small: power decides the force

In electric propulsion, available power sets the thrust. If the efficiency of converting electricity into kinetic energy is η, the kinetic power of the exhaust (½ × flow × exhaust speed²) equals η × P, and since thrust F is flow times exhaust speed, F = 2ηP / (g0 × Isp) Our calculation. For the same power, raising the specific impulse lowers the thrust.

Picking 15 xenon Hall thrusters from NASA's Table 4-10 whose thrust, specific impulse and power are all listed, we get thrust-to-power ratios of about 45 to 66 mN/kW, which corresponds to overall efficiencies of about 0.27 to 0.52. The gridded-ion BIT-3 (iodine) comes to about 14 mN/kW (efficiency about 0.14) Our calculation. These are consistent with NASA's statement of "generally below 75 mN/kW" Sourced[Ref. 1].

4. A map of electric-propulsion types

NASA's report divides electric propulsion by how the propellant is accelerated into electrothermal, electrostatic and electromagnetic types Sourced[Ref. 1]. We turned the thrust and specific-impulse ranges of small-spacecraft devices in NASA's Table 4-1 into a map.

All Sourced (organized from NASA's report [Ref. 1]; the arrangement of the table is Our calculation). The flown examples are the main ones named in that report.

5. Gridded ion thrusters: the "wear" of the grids

A gridded ion thruster turns a gas propellant such as xenon into ions in a plasma discharge and accelerates them backward with a voltage across two grids (the ion optics). The discharge comes in two styles: a direct-current type in which electrons from an internal cathode strike the gas, and a radio-frequency or microwave type with no internal cathode. It runs at high voltage and needs an external cathode, the neutralizer, to keep the beam charge-neutral. High specific impulse is possible, but thrust density is limited by space charge (ions of the same charge repelling each other) Sourced[Ref. 1].

View original SVG, values and sources

Original SVG reference: Concept diagram (vector drawing). Simplified sections of a gridded ion thruster (left) and a Hall thruster (right), and where materials matter. Part names and cautions are from NASA's report [Ref. 1]. The cathode material is from NASA Glenn's inspection report [Ref. 10], and carbon-based grids from NASA's report [Ref. 11].

The grids hold a gap while ions wear them down

NASA's report says the gap between the grids is typically under 1 mm; this narrow gap is what gives the strong field and thrust. The price is that a foreign particle bridging the gap shorts the grids. Charge-exchange ions (ions that picked up an electron on the way and changed character), formed between and behind the grids, strike the grids and slowly erode them. Over time this leads to structural failure of a grid, electrons streaming back, or shorting from deposited conductive material Sourced[Ref. 1].

The spare flight thruster of NASA's Deep Space 1 went through a 30,352-hour extended life test (begun in October 1998 and voluntarily ended on 26 June 2003) Sourced[Ref. 6]. The post-test analysis reported the following Sourced[Ref. 7].

• Enlarged accelerator-grid holes and a narrower grid gap meant that at 26,000 hours the thruster could no longer prevent electron backstreaming at full power
• Pits eroded through the web of the accelerator grid, and grooves reached 45 % of the grid thickness near the center
• Material sputtered from the accelerator grid adhered to the downstream face of the screen grid, but did not spall off as flakes

The long-duration test of NASA Glenn's next-generation thruster NEXT ran from June 2005 to February 2014, reaching 51,184 hours, 918 kg of xenon and a total impulse of 35.5 MN·s before being ended voluntarily Sourced[Ref. 8]. The post-test inspection measured the grid wear in detail Sourced[Ref. 9].

The main points of the NEXT inspection are these Sourced[Ref. 9].

• The change in grid gap was −4 % at the center and −7 % on average. Comparing with −30 % in the NSTAR test, the presentation says the efforts to stabilize the NEXT grid gap were largely successful
• The accelerator grid lost 29.5 g by measurement; correcting for 12.4 g of deposits, mass lost to erosion is estimated at about 42 g (33.7 g in the NSTAR test). The minimum thickness of the screen grid was 86 % of its pre-test value
• Deposits of eroded material and of carbon were found. The carbon is estimated to have arrived by back-sputtering. Deposits reduced the screen-grid hole diameters by 2.2 % on average
• The downstream diameters of the accelerator-grid holes grew by 24 to 33 % (deposits excluded), and the minimum hole diameter by about 5 to 7 %

Carbon-based grids as an alternative

Molybdenum is the conventional grid material, and NASA papers compare carbon-based materials against it Sourced[Ref. 11]. NASA Glenn's report says carbon-based materials have shown nearly an order of magnitude better resistance to sputter erosion than molybdenum. Carbon/carbon composites (carbon fiber in a carbon matrix) and pyrolytic carbon grids have been tried in several programs Sourced[Ref. 11]. A NASA JPL report says 30 cm carbon-carbon grids passed vibration testing undamaged, while also noting that grids of 15 cm and larger have run into many problems, including passing vibration tests Sourced[Ref. 12].

6. Hall thrusters: channel walls and magnetic shielding

NASA's report calls the Hall thruster arguably the most successful electric propulsion by number of units flown. The Soviet Union first flew an SPT-60 in 1971, and over 300 more flew through 2018. The first non-Russian one was the French PPS-1350 on ESA's SMART-1 in 2003. From 2019, 120 Starlink and 6 OneWeb spacecraft each used Hall thrusters, accelerating the count sharply. The report says SpaceX had launched over 11,500 Starlink satellites by April 2026 Sourced[Ref. 1].

The principle: an electrode (anode) sits at the back of an annular discharge channel, with an axial electric field and a radial magnetic field applied. Because the fields cross, electrons circle and advance slowly, and they ionize the neutral propellant, such as xenon, by collision. The ions are accelerated by the electric field and leave backward. Electrons come from a cathode, which also neutralizes the beam. Although its specific impulse is lower than a gridded ion thruster's, a Hall thruster has a higher thrust-to-power ratio, a simple design and good efficiency, which makes it a balanced choice for many missions Sourced[Ref. 1].

The walls erode, and set the life

NASA's report says that the life of a classical Hall thruster is limited mainly by erosion of the discharge channel wall, and that of a magnetically shielded one mainly by erosion of the pole covers. It adds that performance can also change over life through erosion and contamination Sourced[Ref. 1]. The channel walls have been boron nitride (BN) ceramic Sourced[Ref. 13].

NASA JPL showed a magnetic shielding arrangement of the field near the wall that reduces plasma contact with the wall, greatly cuts erosion of the BN walls and extends life by orders of magnitude Sourced[Ref. 13]. Because the wall is now protected, it also became possible to replace the insulating wall with a conductor. In the 6 kW H6 thruster, replacing BN rings with graphite kept efficiency above 60 %, within 2 % of the BN baseline. The graphite walls ran much cooler, and the same report says a thruster without ceramic walls is simpler and cheaper to build to survive launch loads, and radiates more efficiently, which raises the power capability Sourced[Ref. 13].

7. Cathodes and propellants: how materials are chosen

Hollow cathodes: porous tungsten impregnated with barium

Both gridded-ion and Hall thrusters need cathodes to emit electrons. NASA Glenn's inspection report on NEXT examines the cathode insert (the electron-emitting part) as a porous tungsten matrix with an impregnate whose main constituents are barium, oxygen, aluminum and calcium. Using electron microscopy and X-ray diffraction, it looked at impregnate distribution, the thickness of the etched region, and the shape and phase of the emission surface Sourced[Ref. 10]. One expected failure mode is "barium depletion of the insert", which leads to excessive cathode temperature or inability to ignite Sourced[Ref. 10]. In this inspection, the impregnate area fraction of unused inserts and of inserts from the long-duration test agreed within uncertainty, and no significant change in the microstructure of the tungsten matrix and impregnate was seen Sourced[Ref. 10].

NASA's report says cathode life is sensitive to propellant purity and to air and humidity exposure before launch, and that feed-system cleanliness, bake-out and high-purity propellant matter. Heaterless cathodes are also advancing, and JPL's MaSMi work included a 25,000-ignition-cycle heaterless-cathode demonstration Sourced[Ref. 1].

Propellants: what gets thrown out

The "fuel" of electric propulsion is also a materials-compatibility question. NASA's report arranges it as follows Sourced[Ref. 1].

• Xenon: the propellant Hall thrusters have mainly flown on; heavy molecules, low ionization energy, easy to handle
• Krypton and argon: Starlink uses krypton, and its second generation also argon. They are less efficient than xenon and harder to store, but much cheaper; at the scale of thousands of satellites, cost and supply are given as the reasons
• Iodine: three times the storage density of xenon, stored as a solid with low vapor pressure. But it is a strong oxidizer, and its long-term effect on thruster and spacecraft is largely unknown. Beihangkongshi-1, launched in November 2020, made the first on-orbit demonstration of iodine propulsion, with two 90-minute burns changing altitude by 700 m. LunaH-Map, which carried a BIT-3, ended when its propulsion system could not operate; the team concluded that a tank valve was stuck closed (the report does not say whether iodine was involved)
• Water: a propellant for electrothermal (resistojet) systems, used by devices from Bradford, Pale Blue and Aurora
• PTFE and metals: pulsed plasma thrusters ablate PTFE, vacuum-arc thrusters use metals such as molybdenum, and electrosprays use ionic liquids or indium and gallium

8. How long they have run: flight records and life tests

The most important question in judging electric propulsion is "how many hours, and with how much propellant". Here are examples with numbers from NASA sources.

• Dawn (flight): an ion propulsion system developed at NASA JPL. The 1,218 kg spacecraft launched in 2007 and visited Vesta and Ceres. As of June 2016 the propulsion system had run over 48,454 hours, used about 401 kg of xenon and changed speed by over 11.0 km/s, described as a record for an on-board propulsion system Sourced[Ref. 5]
• NEXT long-duration test (ground): 51,184 hours, 918 kg of xenon, 35.5 MN·s Sourced[Ref. 8]
• JPL's MaSMi (ground): a small magnetically shielded Hall thruster. A 7,205-hour wear test used over 100 kg of xenon for a total impulse of about 1.55 MN·s. NASA's table says the same test article is being run on to 8,187 hours and 1.73 MN·s Sourced[Ref. 1]
• BHT-600 (ground): a 7,000-hour test at NASA Glenn ended after 70 kg of xenon Sourced[Ref. 1]
• SPT-70 (ground): experiments have demonstrated a life of 3,100 hours, about 450 kN·s Sourced[Ref. 1]
• GOCE (flight): ESA's gravity satellite (1,000 kg), at an altitude of 220 to 260 km, used two gridded-ion thrusters (T5) to cancel air drag until the xenon ran out in October 2013 Sourced[Ref. 1]
• Hayabusa (flight): JAXA's spacecraft (510 kg, launched on 9 May 2003) made a powered Earth swing-by with its ion engines in May 2004. Demonstrating ion-engine performance was one of its objectives. The title of a paper shows that these were microwave-discharge ion engines Sourced[Ref. 18][Ref. 16][Ref. 17]. JAXA says Hayabusa2's extended mission will use the remaining ion-engine xenon, of which about half remains Sourced[Ref. 19]
• OneWeb (flight): 80 satellites with BHT-350 thrusters launched in December 2022 and January 2023, and as of August 2023 over 100 units were operating, all reported operational Sourced[Ref. 1]

There are also cases that did not go well. LunaH-Map's BIT-3 was inoperable and the mission ended in late 2023 Sourced[Ref. 1]. On UWE-4, three of the four thrusters showed anomalous behavior Sourced[Ref. 1].

9. Effects on the spacecraft: heat and plume

A thruster runs hot right next to the spacecraft and ejects a plasma. The spacecraft's materials feel the effects of this "neighbor".

View original SVG, values and sources

Original SVG reference: Concept diagram (vector drawing). What a thruster does to its spacecraft. Heat soak-back and plume effects are from NASA's report [Ref. 1] and the abstract of the SPT-100 witness-plate test [Ref. 15].

• Heat: most of a Hall thruster's heat is radiated directly from its surfaces, but some soaks back to the spacecraft through the mount, lines and harness. Low-efficiency types such as ambipolar and pulsed thrusters carry a larger heat load Sourced[Ref. 1]. For the thermal design of the body see our explainer on spacecraft thermal control
• Plume (the exhaust plasma): plume ions of a Hall thruster erode or contaminate surfaces even at large angles. Ground measurements at large angles may under-predict what happens in flight Sourced[Ref. 1]
• A witness-plate example: NASA exposed witness plates to the plume of an SPT-100 for 50 hours. In the most energetic region there was a net removal of material from samples facing the thruster, and at the periphery net deposits were seen along with changes in optical properties. Samples inside collimators changed within experimental uncertainty Sourced[Ref. 15]
• Electrosprays: the propellant is conductive, and deposits can short circuits or foul solar panels. Plume shields are often used to protect surfaces Sourced[Ref. 1]

10. Limits, and what could not be confirmed

• Scope of this article: it rests on NASA's report on small-spacecraft devices (spacecraft of 180 kg or less) and on NASA technical reports. We did not check primary sources on high-power Hall thrusters for large geostationary satellites, nuclear-electric propulsion or magnetoplasma propulsion, so we do not cover them
• Propellant price and supply: we do not use market estimates, so we give no price figure for xenon
• Sources checked by abstract only: many of the NTRS reports behind Figs. 6 to 8 were checked by abstract only (stated in the references). The NEXT inspection presentation and the insert report were checked in full text
• Hayabusa numbers: we could not confirm operating hours, propellant mass or performance figures for its ion engines from JAXA's official pages or the NASA abstract. We checked only the title and bibliographic data of the IEEE paper, not its text
• Grid materials: we do not state the grid material of any specific engine where the sources did not confirm it (molybdenum or carbon-based)
• Calculations: the propellant fractions, thrust-to-power ratios and days are illustrative values from simple formulas, not predictions for any real spacecraft
• The text of standards: we did not check the requirements in the text of the ECSS standards. We did not find a published ISO standard on electric propulsion within the range we checked
• ASTM: we did not open ASTM's site

11. How to read the yardsticks and standards

Electric propulsion is evaluated by combining "performance measured in a ground vacuum chamber", "flight record" and "inspection of material wear". The yardsticks are: (1) thrust, specific impulse and efficiency (at nominal operating conditions); (2) total impulse and propellant throughput (kg); (3) operating hours in life tests and the wear measured afterward; (4) repeated cycles such as ignition counts; (5) the effect of the plume on other surfaces; and (6) development maturity. All of them depend on the conditions of the test facility.

The arrangement of this table is Our calculation (our own reordering). Standard numbers, titles and status were checked on the official pages of ECSS and ISO (checked 17 October 2026). ISO texts are paywalled, so we limited ourselves to titles and scope. We did not open ASTM's site.

Cautions in reading (for materials engineers)

• "Nominal" is not "full range": NASA's table lists values only at each device's nominal operating condition, and asks readers to follow the references for the full range
• Ground tests and flight are different things: facility back-pressure and the angle of plume measurements can make ground values differ from flight
• Some numbers are unverified: NASA's table carries a note not to assume data has been independently verified unless otherwise published
• Check whether a standard is published: something under development, like ISO/DTR 26097, cannot be used as a requirement

12. Glossary

13. References

1. NASA (Small Spacecraft Systems Virtual Institute) "State-of-the-Art of Small Spacecraft Technology", Chapter 4 In-Space Propulsion (2026 edition, including Tables 4-1, 4-9, 4-10 and 4-11; checked 17 October 2026) — nasa.gov/smallsat-institute/sst-soa/in-space-propulsion/
2. NASA Glenn Research Center "Specific Impulse" (definitions of thrust, total impulse and specific impulse Isp = Veq/g0; checked 17 October 2026) — grc.nasa.gov/www/k-12/airplane/specimp.html
3. NASA Glenn Research Center "Ideal Rocket Equation" (checked 17 October 2026) — www1.grc.nasa.gov/beginners-guide-to-aeronautics/ideal-rocket-equation/
4. NIST "standard acceleration of gravity" (9.806 65 m/s2, exact; CODATA 2022; checked 17 October 2026) — physics.nist.gov/cgi-bin/cuu/Value?gn
5. Garner C. E., Rayman M. D. (NASA JPL) "In-flight Operation of the Dawn Ion Propulsion System Through the Low Altitude Mapping Orbit at Ceres" (2016; abstract only) — ntrs.nasa.gov/citations/20190025608
6. Sengupta A., Brophy J. R., Goodfellow K. D. (NASA JPL) "Status of the extended life test of the Deep Space 1 flight spare ion engine after 30,352 hours of operation" (abstract only) — ntrs.nasa.gov/citations/20060043359
7. Anderson J. R., Sengupta A., Brophy J. R. (NASA JPL) "Post-Test Analysis of the Deep Space One Spare Flight Thruster Ion Optics" (abstract only) — ntrs.nasa.gov/citations/20110013157
8. Shastry R., Herman D. A., Soulas G. C., Patterson M. J. (NASA Glenn) "End-of-Test Performance and Wear Characterization of NASA's Evolutionary Xenon Thruster (NEXT) Long-Duration Test" (2014; abstract only) — ntrs.nasa.gov/citations/20140016766
9. Soulas G. C., Shastry R. (NASA Glenn) "Post-Test Inspection of NASA's Evolutionary Xenon Thruster Long Duration Test Hardware: Ion Optics" (presentation of 25 July 2016; full text checked) — ntrs.nasa.gov/citations/20170002704
10. Mackey J., Shastry R., Soulas G. (NASA Glenn) "Characterization of the NEXT Hollow Cathode Inserts After Long-Duration Testing" (text checked) — ntrs.nasa.gov/citations/20170009085
11. Haag T., Patterson M., Rawlin V., Soulas G. (NASA Glenn) "Carbon-Based Ion Optics Development at NASA GRC" (2002; abstract only) — ntrs.nasa.gov/citations/20020061976
12. Snyder J. S., Brophy J. R. (NASA JPL) "Performance Characterization and Vibration Testing of 30-cm Carbon-Carbon Ion Optics" (abstract only) — ntrs.nasa.gov/citations/20070022514
13. Goebel D. M., Hofer R. R., Mikellides I. G., Katz I., Polk J. E., Dotson B. (NASA JPL) "Conducting Wall Hall Thrusters" (2013; abstract only) — ntrs.nasa.gov/citations/20150007838
14. Mikellides I. G., Ortega A. L. (NASA JPL) "Assessment of Pole Erosion in a Magnetically Shielded Hall Thruster" (2014; abstract only) — ntrs.nasa.gov/citations/20160008257
15. Pencil E. J. (NASA Lewis) "Preliminary far-field plume sputtering characterization of the Stationary Plasma Thruster (SPT-100)" (abstract only) — ntrs.nasa.gov/citations/19940020293
16. Abell P. A. (NASA) "The Hayabusa Spacecraft Sample Return Mission" (abstract only; launched 9 May 2003; demonstrating ion-engine performance was one objective) — ntrs.nasa.gov/citations/20100033487
17. Kuninaka H., Nishiyama K., Funaki I., Shimizu Y., Yamada T., Kawaguchi J. "Assessment of Plasma Interactions and Flight Status of the HAYABUSA Asteroid Explorer Propelled by Microwave Discharge Ion Engines", IEEE Transactions on Plasma Science, 34(5), 2125-2132 (2006; title and bibliographic data only; text not checked) — doi.org/10.1109/TPS.2006.876513
18. JAXA Institute of Space and Astronautical Science (ISAS) "HAYABUSA" (mass 510 kg, launched on M-V-5, powered Earth swing-by with ion engines in May 2004; checked 17 October 2026) — isas.jaxa.jp/en/missions/spacecraft/past/hayabusa.html
19. JAXA Institute of Space and Astronautical Science (ISAS) "Hayabusa2" (the extended mission uses the remaining ion-engine xenon, about half; checked 17 October 2026) — isas.jaxa.jp/en/missions/spacecraft/current/hayabusa2.html
20. ECSS ECSS-E-ST-35C Rev.1 "Propulsion general requirements" (6 March 2009; checked 17 October 2026) — ecss.nl/standard/ecss-e-st-35c-rev-1-propulsion-general-requirements/
21. ECSS ECSS-E-ST-35-01C "Liquid and electric propulsion for spacecraft" (15 November 2008; checked 17 October 2026) — ecss.nl/standard/ecss-e-st-35-01c-liquid-and-electric-propulsion-for-spacecraft/
22. ISO ISO/DTR 26097 "Space systems — Micro-cathode arc propulsion module for CubeSats" (draft technical report; status "under development", stage 50.00; checked 17 October 2026) — iso.org/standard/92605.html

14. Claim-to-source audit

Last updated: 17 October 2026 / Sources are limited to primary materials (NASA's report on small-spacecraft technology, NASA technical reports, NASA Glenn, NIST, JAXA, the official pages of ECSS and ISO, and bibliographic data of peer-reviewed papers). The propellant fractions, thrust-to-power ratios and days are illustrative calculations made for this article and do not predict any real spacecraft. We could not confirm high-power Hall thrusters, propellant prices, the grid material of each engine, numbers for Hayabusa's ion engines, or the requirements in the standards' texts, so we do not state them.