Gridded Ion Thruster

Two metal grids separated by a millimetre extract xenon ions from a plasma and hurl them at 40 kilometres per second — the most fuel-efficient chemical-free engine ever flown in deep space.

A gridded ion thruster converts electrical power — from solar panels or a nuclear source — into the kinetic energy of a xenon ion beam. Inside the discharge chamber, electrons emitted by a hollow cathode ionise xenon propellant by electron-impact ionisation, creating a xenon plasma. Positive Xe⁺ ions drift toward two closely spaced grids at the downstream end of the chamber. The screen grid (held at high positive voltage) allows ions to pass through its perforations; the accelerator grid (held at negative voltage) accelerates the ions to exhaust velocities of 30–80 km/s. A separate neutraliser hollow cathode emits electrons into the beam to prevent spacecraft charge buildup. The result is a thrust of 25–235 mN (NEXT) consuming 0.5–6.9 kW of power, with a total efficiency of 50–70% and a specific impulse (Isp) of 2800–4190 s — versus 300–450 s for a chemical bipropellant engine. The low thrust demands long operating times (months to years), so gridded ion thrusters are designed for operational lifetimes exceeding 30,000 hours.

Xenon is chosen as propellant because its ionisation potential (12.13 eV) is relatively low, its atomic mass (131.3 u) is high enough to produce good momentum per ion, it is chemically inert and non-toxic, and it can be stored as a high-pressure gas without cryogenic handling. In a Kaufman-type direct current discharge chamber, the hollow cathode injects electrons that spiral in the magnetic field created by permanent magnets around the chamber walls, maximising collision probability with xenon atoms before reaching the anode. The ion extraction grid system operates on a straightforward electrostatic principle: the screen grid voltage (V_s, typically +1000 to +1800 V) sets the ion acceleration energy; the accelerator grid voltage (V_a, typically −200 to −500 V) prevents back-streaming of electrons from the neutraliser. The grid gap is approximately 1 mm — tight enough to form a debye sheath that collimates the ion beam but not so tight that sputtering erosion destroys the grid in less than an operational lifetime. The exhaust velocity v_e = √(2qV_s/m) scales with the square root of the beam voltage, so doubling voltage from 1100 V (NSTAR) to 1800 V (NEXT) increases Isp from ~3100 s to ~4190 s at full power. Thruster efficiency η = T²/(2 P ṁ_p), where T is thrust, P is input power, and ṁ_p is propellant mass flow rate.

In plain terms

Imagine a supersoaker that fires invisible bullets instead of water. But instead of using gas pressure, it uses electricity to push the bullets out faster than a bullet train — so fast that it barely uses any ‘bullets’ to push the spacecraft forward. That’s what an ion thruster does with xenon gas atoms.

Also called: electrostatic ion thruster, Kaufman thruster, gridded electrostatic thruster, ion engine, NSTAR, NEXT thruster.

Why it matters

Specific impulse of 2800–4190 s versus 300–450 s for hydrazine monopropellant means the same mass of propellant produces 7–14× more velocity change (Δv), allowing missions — like Dawn's orbit of two separate asteroid-belt bodies — that are impossible with chemical propulsion given a reasonable launch mass.
The NSTAR thruster on Deep Space 1 (1998) was the first ion drive used as primary propulsion on a NASA deep-space mission, demonstrating reliability over 16,265 hours of in-space operation and validating the technology for all subsequent ion-propelled missions.
GOCE (ESA, 2009–2013) used QinetiQ T5 Kaufman thrusters in drag-compensation mode, throttling continuously to counteract aerodynamic drag in a 254 km orbit — enabling gravity-gradient measurements 100× more precise than any predecessor.
Grid erosion by sputtering is the primary lifetime limiter; NEXT's NSTAR-heritage carbon-carbon composite grids (replacing molybdenum) demonstrated >48,000 hours in life tests, enabling missions lasting more than five years at low thrust.

Where it lives

Deep Space 1 (NASA, 1998–2001) — single NSTAR thruster, first NASA deep-space ion drive primary propulsion; 16,265 hours in-flight operation, 74 kg xenon consumed
Dawn (NASA, 2007–2018) — three NSTAR 30-cm thrusters; orbited asteroid Vesta and dwarf planet Ceres, the only spacecraft to orbit two extraterrestrial bodies
Hayabusa and Hayabusa2 (JAXA, 2003/2014) — μ10 ECR microwave-discharge ion engines; asteroid sample return from Itokawa and Ryugu respectively
BepiColombo (ESA/JAXA, 2018–present) — four QinetiQ T6 gridded ion engines; primary propulsion for the journey to Mercury, 22 kN·s total impulse design
GOCE (ESA, 2009–2013) — QinetiQ T5 thrusters in drag-free orbit at 254 km altitude, throttling 0–20 mN to counteract atmospheric drag continuously
Psyche (NASA, 2023–present) — four Hall thrusters for primary propulsion (Psyche uses Hall, not gridded ion — gridded ion missions noted above are all confirmed)

Key numbers

Specific impulse (Isp) — NEXT at full power: 4,190 s (Patterson & Benson 2007; NSTAR at full throttle 3,100 s; chemical bipropellant 300–450 s)
Maximum thrust — NEXT: 235 mN (At 6.9 kW input power and 1800 V beam voltage)
Beam voltage: NSTAR: 1,100 V; NEXT: 1,800 V (Directly sets exhaust velocity: v_e = √(2eV_b/m_Xe); higher voltage → higher Isp but more power per unit thrust)
Total efficiency: 50–70% (Including discharge losses, neutraliser power, and power processing unit losses; best achieved by NEXT at high-throttle points)
Xenon ionisation potential: 12.13 eV (Low enough for efficient electron-bombardment ionisation; first ionisation potential from NIST ASD)
Grid gap: ~1 mm (Screen-to-accelerator grid spacing; sets Debye sheath structure and determines grid erosion lifetime)
Demonstrated operational life (NEXT): >48,000 hours (Extended Life Test at Glenn Research Center; grid erosion was primary lifetime limiter)

Surprising facts

The NEXT thruster operating at 6.9 kW produces a maximum thrust of 235 mN — approximately the weight of a large apple. Yet applied continuously over years, this force can change a spacecraft's heliocentric velocity by kilometres per second, making orbit changes around two different asteroid-belt bodies feasible from a single launch.
Xenon ions exit the NEXT accelerator grid at approximately 80 km/s at full beam voltage (1800 V). The Space Shuttle main engine's hydrogen exhaust velocity was about 4.4 km/s. The gridded ion thruster's exhaust is 18× faster.
The accelerator grid in a high-power ion thruster is perforated by thousands of small holes (a few millimetres diameter, precisely aligned with the screen grid holes above them) and operates at a voltage several hundred volts below cathode potential. A misalignment error of even tens of microns would cause the ion beam to strike the accelerator grid, eroding it in hours. Grid manufacturing tolerance is held to ±25 µm across the full 30 cm aperture.
Harold Kaufman demonstrated the first practical electron-bombardment ion engine at NASA Lewis Research Center in 1959, using mercury as propellant. Mercury was abandoned in the 1980s because of contamination of spacecraft surfaces and handling hazards, and xenon became the propellant of choice from the mid-1980s onward.

Open questions

Grid erosion pattern is self-sustaining and hard to predict: Ion thruster accelerator grids erode by sputtering: Xe⁺ ions that are not perfectly collimated strike the grid material and knock out atoms. The erosion creates dish-shaped pits around each aperture. As the geometry changes, the electric field distribution near each aperture changes, which alters the ion trajectory distribution, which changes the erosion pattern. This feedback between erosion geometry and ion optics makes long-term erosion prediction a coupled simulation problem that still cannot be reliably done from first principles without empirical calibration. End-of-life grid failure has occurred on flight hardware in ways that ground life tests did not fully predict.

A short history

The concept of electric propulsion — accelerating ions electrically rather than thermally — dates to Konstantin Tsiolkovsky (who discussed it in 1911) and Robert Goddard (who patented ion acceleration in 1917). Ernst Stuhlinger at the Army Ballistic Missile Agency wrote the first detailed electric propulsion mission analysis in the 1950s. Harold Kaufman at NASA's Lewis Research Center (now Glenn) built the first practical electron-bombardment ion engine in 1959, using mercury propellant. The SERT-1 spacecraft (1964) was the first in-space test of an ion engine, and SERT-II (1970) demonstrated long-duration space operation. Through the 1970s–80s, Soviet/Russian spacecraft flew Hall-effect thrusters extensively, but Western use of gridded ion engines was limited to station-keeping: Hughes (later Boeing) flew the XIPS-13 and XIPS-25 on commercial GEO satellites from the early 1990s. The switch to xenon propellant from mercury occurred in the 1980s due to contamination and handling concerns. NASA JPL's NSTAR programme, developed for the New Millennium Program, flew on Deep Space 1 (launched 1998) — the first deep-space mission with an ion drive as primary propulsion. The success of Deep Space 1 and Dawn (2007) established gridded ion thrusters as the standard for high-Δv deep-space science missions. ESA's BepiColombo (2018) carries the highest-thrust gridded ion system flown to date for a planetary mission, using four T6 thrusters rated at 230 mN each.

Sources

Goebel D.M. & Katz I. — 'Fundamentals of Electric Propulsion: Ion and Hall Thrusters' (JPL Space Science and Technology Series / Wiley, 2008). Free PDF via JPL (https://descanso.jpl.nasa.gov/SciTechBook/series1/Goebel__cmprsd_opt.pdf)
Patterson M.J. & Benson S.W. — 'NEXT Ion Propulsion System Development Status and Performance', AIAA-2007-5199 (2007) (https://ntrs.nasa.gov/citations/20070032696)
Polk J.E. et al. — 'Validation of the NSTAR Ion Propulsion System on the Deep Space One Mission', AIAA-99-2446 (1999) (https://ntrs.nasa.gov/citations/19990064058)
Rayman M.D. et al. — 'Dawn: A mission in development for exploration of main belt asteroids Vesta and Ceres', Acta Astronautica 58(11), 605–616 (2006). DOI: 10.1016/j.actaastro.2005.12.001 (https://doi.org/10.1016/j.actaastro.2005.12.001)
NIST Atomic Spectra Database — Xenon ionisation energies (https://physics.nist.gov/PhysRefData/ASD/ionEnergy.html)