Chemical thrusters burn propellant for a strong, short push; electric thrusters use electricity to fling out gas very fast, giving a tiny push but using far less propellant.
Both kinds of engine work the same way at heart: throw mass backwards, move forwards. The difference is where the energy comes from.
A chemical thruster gets its energy from the propellant itself, by burning fuel with an oxidiser or by breaking down a single chemical such as hydrazine. An electric thruster gets its energy from the solar panels and uses it to accelerate charged gas.
Chemical thrusters push hard. NASA lists hydrazine thrusters for small spacecraft at up to about 28 newtons. Electric thrusters push gently: NASA's figures for Hall-effect thrusters are 0.25 to 55 millinewtons.
But electric thrusters squeeze far more out of each kilogram. Their exhaust is much faster, which engineers measure as specific impulse. NASA gives hydrazine thrusters 200–235 seconds; Hall-effect thrusters reach up to about 1,920 seconds, and gridded ion engines up to 3,000.
ISRO first flew electric propulsion on GSAT-9, the South Asia Satellite, launched in May 2017. ISRO said GSAT-9 used conventional chemical bipropellant thrusters together with electric propulsion, the first time it used electric propulsion for station-keeping.
In March 2025 ISRO finished a 1,000-hour life test of a 300 millinewton Stationary Plasma Thruster running at 5.4 kW. It uses xenon and, according to ISRO, has a specific impulse at least six times that of conventional propulsion. ISRO plans to fly it on its Technology Demonstration Satellite (TDS-01) to raise an orbit all the way to geostationary.
Less propellant means a lighter satellite, or room for more transponders — the reason ISRO gives for moving to electric propulsion. The price is time. A gentle push takes weeks or months to do what a chemical engine does in hours.
Worked example (illustrative): a 2,000 kg satellite needs to change its speed by 1,000 m/s over its life. With hydrazine at 230 s it needs about 716 kg of propellant. With a Hall-effect thruster at 1,800 s it needs only about 110 kg. But ISRO's 300 mN thruster would need about 77 days of firing to deliver that change (see the calculation below).
Find GSAT-9 (South Asia Satellite) among the geostationary satellites: it is India's first satellite to hold its slot partly with plasma thrusters. Open the ISRO satellite tracker
The rocket equation shows why exhaust speed matters so much:
m_propellant = m_start × [1 − e^(−Δv ÷ v_e)], v_e = Isp × g₀, g₀ = 9.81 m/s²
Worked example (illustrative): a 2,000 kg satellite needs Δv = 1,000 m/s over its life.
The electric option saves about 600 kg — mass that can become payload.
Now the cost in time. With ISRO's 300 mN thruster on the same 2,000 kg satellite:
a = F ÷ m = 0.3 ÷ 2,000 = 0.00015 m/s²
In one day (86,400 s) that gives 0.00015 × 86,400 ≈ 13 m/s. The full 1,000 m/s takes about 77 days of firing. A chemical engine delivers it in minutes.
Power sets the limit. NASA notes electric propulsion generally gives below about 75 mN per kW; ISRO's thruster gives 300 mN at 5.4 kW, about 56 mN/kW. So a satellite's solar array must be sized for the thruster as well as the payload. Thrust rises with power; for a fixed power, higher exhaust speed means less thrust, since jet power ≈ ½ × F × v_e.
They expel propellant at a much higher speed (higher specific impulse), so each kilogram of propellant gives a much larger change in velocity.
GSAT-9, the South Asia Satellite, launched in May 2017; its plasma thrusters assisted station-keeping.
Very low thrust — millinewtons — so manoeuvres take weeks or months, and the thrusters need kilowatts of electrical power.