Chapter 20: Electric Propulsion

"Slow and steady wins the race." — the moral of Aesop's fable "The Tortoise and the Hare"

Overview

For nineteen chapters the rocket equation has been a tyrant. It set the tempo of everything: because the exhaust velocity of a chemical rocket is stuck near $4.5\ \text{km/s}$, and because velocity change grows only with the logarithm of the mass ratio, we are forced into rockets that are ninety percent propellant, into staging, into a relentless war on structural mass. Chapter 3 called this the tyranny of the rocket equation, and every propulsion chapter since has been a response to it — better propellants (Chapter 18), better nozzles (Chapter 19), better engine cycles (Chapter 17). All of them nibble at the edges. None of them changes the fundamental fact that a chemical engine can throw its exhaust only so fast, because the energy comes from the chemistry of the propellant itself, and chemistry has a ceiling.

This chapter is about a different idea entirely — one that walks up to the tyrant and changes the rules. What if the energy that accelerates the propellant did not come from the propellant? What if you carried a tank of inert gas as pure reaction mass, and a separate supply of electricity — from solar panels, or one day a reactor — and used that electricity to hurl the gas out the back at ten times the speed any chemical reaction could manage? You would have broken the chemical ceiling. Your specific impulse would leap from a few hundred seconds to a few thousand. The rocket equation would suddenly be on your side: the same mission that demanded a rocket that is 80% propellant would now need one that is 15% propellant.

There is, of course, a catch, and it is a big one. The thrust of such an engine is measured not in meganewtons, not in kilonewtons, but in milli-newtons — roughly the weight of a coin resting on your palm. You could no more launch with it than you could drive a car by blowing on it. But in space, where you have all the time in the world and nothing to push against anyway, a whisper of thrust applied for months or years adds up to a hurricane of delta-v. This is electric propulsion, and it has quietly become the workhorse of the modern satellite and the enabler of missions — orbiting two asteroids on one tank, flying an armada of ten thousand satellites — that chemistry alone could never afford.

In this chapter, you will learn to:

• Explain how electric propulsion decouples the energy source from the reaction mass, and why that lifts the chemical ceiling on exhaust velocity.
• Describe how gridded ion engines and Hall thrusters turn electricity and xenon into thrust, and compare them.
• Use the power–thrust–Isp relationship to compute a real thruster's thrust and propellant flow.
• See, in hard numbers, why an ion engine can cross the solar system but cannot lift itself off a table.
• Compute a spiral transfer's delta-v, compare it with a Hohmann transfer, and decide when the trade is worth it.

Learning Paths

🚀 Space Enthusiast: Read 20.1 for the big idea (the tortoise beats the hare), skim the grid physics in 20.2, and spend your time on 20.5 (why it cannot launch) and 20.6 (Dawn, Starlink, BepiColombo). The decoupling idea in 20.1 is the one that will change how you think about engines.

📐 Engineering Student: Read everything. The power–thrust–Isp derivation in 20.1 and 20.5, the ion-acceleration formula in 20.2, and the spiral-transfer comparison in 20.6 are the quantitative core; do the ⭐⭐/⭐⭐⭐ exercises. This chapter leans on Chapters 3 and 16 constantly — keep them open.

🎮 KSP Player: Ion engines in the game feel exactly like the real thing: tiny thrust, absurd delta-v, burns that last "physics-warp" minutes. Focus on 20.5 and 20.6 to understand why your ion probe takes forever to change orbit but sips fuel. The spiral-vs-Hohmann section explains the long, patient burns.

🛰️ Industry Prep: Electric propulsion is now standard on comsats and constellations. Sections 20.3 (Hall thrusters — the commercial workhorse), 20.5 (the power–thrust trade), and 20.6 (all-electric orbit-raising and station-keeping) are directly the language of modern satellite propulsion. Your Mission Design Checkpoint decides whether your mission should fly electric.

20.1 The high-Isp, low-thrust idea

Start with the one equation that governs every rocket, chemical or electric — the thrust equation you met rigorously in Chapter 16. A rocket's thrust is the rate at which it throws momentum out the back:

$$ F = \dot m\, v_e, $$

where $\dot m$ is the mass flow rate of propellant (kg/s) and $v_e$ is the effective exhaust velocity (m/s). Two engines can produce the same thrust in completely different ways: one can throw a lot of mass slowly, and the other can throw a little mass very fast. Chemical rockets live at the first extreme. An ion engine lives at the second. That single choice — where you sit on the trade between $\dot m$ and $v_e$ — is the whole story of this chapter.

Definition (electric propulsion). Electric propulsion is any rocket that uses electrical energy — rather than the chemical energy of combustion — to accelerate its propellant to high exhaust velocity. The propellant (typically an inert gas) is merely reaction mass; the energy comes from a separate source, usually solar arrays or, in principle, a nuclear reactor.

The word separate is doing enormous work in that definition, and it is worth dwelling on, because it is the physical heart of everything that follows.

In a chemical rocket, the propellant plays two roles at once: it is both the energy source and the reaction mass. The kerosene and oxygen you burn are the energy (their chemical bonds rearrange and release heat) and they are the stuff you throw out the back. Bundling those two jobs into one substance is convenient — it is why chemical rockets are so simple and so powerful — but it is also a prison. The exhaust velocity you can reach is set by how much energy a kilogram of propellant can release, and no chemical reaction we know releases enough to push exhaust much past $4.5\ \text{km/s}$. You cannot buy a higher $v_e$ at any price, because the currency — chemical bond energy — is capped by nature.

Electric propulsion breaks the two roles apart. The reaction mass is now an inert gas like xenon, which stores no useful chemical energy at all; it is just heavy atoms to throw. The energy comes from somewhere else entirely — sunlight on a solar array, converted to electricity, delivered to the thruster by wire. And here is the liberation: once energy and reaction mass are separate, there is no longer any chemical ceiling on exhaust velocity. You can pour as much energy into each kilogram of xenon as your power supply can deliver. If you have the watts, you can accelerate that xenon to $30$, $40$, even $50\ \text{km/s}$ — ten times what the best chemical engine can do.

🚪 Threshold Concept: energy and propellant are separate things. In a chemical rocket, propellant is energy. This is so deeply built into our intuition that it takes a moment to see past it. Electric propulsion severs the link: the propellant is what you throw (xenon), and the energy is what accelerates it (electricity). Once you internalize this split, the whole logic of the field falls out. Specific impulse is no longer bounded by chemistry, so it can be enormous — that is the good news. But the power to accelerate the propellant must now be generated, carried, and paid for in mass and watts — that is the catch, and the source of the low thrust. Every property of an electric thruster, good and bad, traces back to this one act of decoupling. It changes how you see every engine after it.

Let us make the payoff concrete with the tool from Chapter 3. Recall the rocket equation, $\Delta v = v_e \ln(m_0/m_f)$, and its inverse, the mass ratio $m_0/m_f = e^{\Delta v/v_e}$. Because $v_e$ sits in the denominator of that exponent, a tenfold increase in $v_e$ shrinks the exponent tenfold — and the mass ratio, and therefore the propellant, collapses.

Worked Example: the same maneuver, chemical vs. electric. A deep-space probe with a dry mass of $m_f = 1{,}000\ \text{kg}$ must perform a $\Delta v = 5\ \text{km/s}$ maneuver.

With a chemical engine ($I_{sp} = 320\ \text{s}$, so $v_e = 320 \times 9.81 = 3{,}138\ \text{m/s}$): $$\frac{m_0}{m_f} = e^{5000/3138} = e^{1.593} = 4.92, \qquad m_p = 1000\,(4.92 - 1) = 3{,}920\ \text{kg}.$$

With an electric engine ($I_{sp} = 3{,}000\ \text{s}$, so $v_e = 3{,}000 \times 9.81 = 29{,}420\ \text{m/s}$): $$\frac{m_0}{m_f} = e^{5000/29420} = e^{0.170} = 1.19, \qquad m_p = 1000\,(1.19 - 1) = 185\ \text{kg}.$$

The chemical stage needs $3{,}920\ \text{kg}$ of propellant; the electric stage needs $185\ \text{kg}$ — twenty-one times less, for exactly the same job. That saved $3{,}700\ \text{kg}$ is the entire reason electric propulsion exists. It is the rocket equation, for once, working for you instead of against you.

That is the high-Isp half of the idea, and it is glorious. Now the low-thrust half, which we will spend the rest of the chapter reckoning with. The catch hides inside the word power. The kinetic power carried away in the exhaust beam — the "jet power" — is

$$ P_{\text{jet}} = \tfrac{1}{2}\,\dot m\, v_e^2. $$

Combine this with $F = \dot m\, v_e$ by eliminating $\dot m$, and you get a relationship that will haunt the rest of the chapter:

$$ \boxed{\ F = \frac{2\,P_{\text{jet}}}{v_e} = \frac{2\,\eta\,P}{v_e}\ } $$

where $P$ is the electrical power drawn from the spacecraft and $\eta = P_{\text{jet}}/P$ is the thruster's total efficiency (how much of the electrical power ends up as directed kinetic energy in the beam; for real thrusters $\eta$ is roughly $0.5$ to $0.7$). Read this equation slowly, because it is the one that explains everything strange about electric propulsion. For a fixed amount of power, thrust is inversely proportional to exhaust velocity. The very thing that makes electric propulsion efficient — a huge $v_e$ — is the thing that makes its thrust tiny. You cannot have both high $v_e$ and high thrust without pouring in more power, and power, as we will see, is heavy and scarce in space.

💡 Intuition: Think of a fixed power budget as a fixed number of dollars per second to spend on momentum. Momentum per unit energy is $F/P_\text{jet} = 2/v_e$: the faster you insist on throwing each kilogram, the less momentum you buy per joule, because kinetic energy grows as $v_e^2$ while momentum grows only as $v_e$. High exhaust velocity is expensive momentum. That is why an ion engine, sipping a couple of kilowatts, produces a thrust you could balance on a fingertip.

🔄 Check Your Understanding 1. In one sentence, what does electric propulsion "decouple" that a chemical rocket bundles together, and why does that lift the ceiling on $v_e$? 2. Using $F = 2\eta P/v_e$, if you double a thruster's exhaust velocity while keeping its power and efficiency fixed, what happens to its thrust?

Answers

1. It decouples the energy source (electricity) from the reaction mass (inert propellant); because the energy no longer comes from the propellant's own chemistry, there is no chemical limit on how fast you can accelerate the propellant — only a power limit. 2. Thrust halves. With $P$ and $\eta$ fixed, $F \propto 1/v_e$, so doubling $v_e$ cuts thrust in half (and, from $\dot m = F/v_e$, cuts the mass flow to a quarter).

20.2 Ion engines: accelerating xenon through a grid

The most efficient electric thrusters flown are gridded ion engines, and they are also the easiest to understand, because they do exactly one thing: they take charged atoms and pull them through a voltage.

Definition (ion engine). A gridded electrostatic ion engine ionizes a propellant gas, then accelerates the positive ions to high velocity by pulling them through a strong electric field established between a pair of closely spaced perforated electrodes (the grids). A separate cathode sprays electrons into the departing beam to keep it electrically neutral.

Follow a single atom of propellant through the engine. It begins as a neutral atom of xenon drifting into a discharge chamber. There it is struck by an energetic electron (from a hollow cathode, or in some designs from radio-frequency fields) hard enough to knock off one of its own electrons. The atom is now a positive ion, $\text{Xe}^+$, and — crucially — being charged, it can be pushed by an electric field. At the downstream end of the chamber sit two grids a millimeter or so apart: a screen grid held at a high positive potential (say $+1{,}000\ \text{V}$) and, just behind it, an accelerator grid held negative. The ion feels this voltage drop as a cliff and is flung through the aligned holes in the grids, leaving the engine at tens of kilometers per second.

How fast, exactly? An ion of charge $q$ and mass $m$ that falls through a potential difference $V$ gains kinetic energy equal to the electrical work done on it, $qV = \tfrac{1}{2} m v_e^2$. Solve for the exit speed:

$$ v_e = \sqrt{\frac{2\,q\,V}{m}}. $$

This is the ion engine's version of the exhaust-velocity story — and notice it contains no chemistry at all, only the voltage you choose and the ion you pick. Let us put in xenon's numbers.

Worked Example: how fast does xenon leave? A xenon ion has charge $q = 1.602\times10^{-19}\ \text{C}$ (one missing electron) and mass $m = 131.3\ \text{u} = 131.3 \times 1.6605\times10^{-27} = 2.18\times10^{-25}\ \text{kg}$. Accelerate it through $V = 1{,}000\ \text{V}$: $$v_e = \sqrt{\frac{2\,(1.602\times10^{-19})(1{,}000)}{2.18\times10^{-25}}} = \sqrt{1.47\times10^{9}} = 3.83\times10^{4}\ \text{m/s} \approx 38\ \text{km/s}.$$ That is an exhaust velocity of $38\ \text{km/s}$ — a specific impulse of $v_e/g_0 = 38{,}300/9.81 \approx 3{,}900\ \text{s}$, roughly ten times the best chemical engine, from a single kilovolt. Sanity check: real ion engines quote $I_{sp}$ nearer $3{,}000$–$3{,}500\ \text{s}$, a little below our ideal figure — as they should be, because not every ion sees the full voltage, some propellant escapes without being ionized, and the beam spreads slightly. The formula gives the ceiling; reality sits just under it.

Two design consequences fall straight out of this picture, and they define the ion engine's character.

Why the thrust is so small: the space-charge limit. The ions are all positive, and like charges repel. As you try to cram more ion current through the gap between the grids, the cloud of ions in flight builds up its own repulsive electric field that chokes off the flow — a phenomenon called space-charge limiting (it obeys the Child–Langmuir law, if you want the name). In practice this caps the ion current, and therefore the mass flow $\dot m$, at a small value for a given grid area and voltage. Small $\dot m$ times a large $v_e$ gives, via $F = \dot m v_e$, only a small thrust. This is not an engineering failure to be fixed; it is a fundamental limit of pulling charged particles through a gap. It is the reason gridded ion engines top out at tens or a couple hundred milli-newtons.

Why you must neutralize the beam. The engine throws out a stream of positive ions and keeps their electrons behind. If you did nothing, the spacecraft would rapidly charge up negative, and — opposite charges attracting — it would simply reel the departing ions back in. No net momentum would leave, and you would get no thrust at all. The fix is a neutralizer: a cathode near the exit that sprays electrons into the beam at the same rate ions leave, so the exhaust (and the spacecraft) stays neutral. It is easy to overlook, but the neutralizer is not optional plumbing — without it the engine does not work as a rocket at all.

📜 From History: Ion propulsion is older than it looks. Robert Goddard sketched electrostatic ion acceleration in his notebooks around 1906, and Konstantin Tsiolkovsky speculated about electric thrust not long after — both men, once again, decades ahead of the hardware. The first ion engine actually operated in space was NASA's SERT-1 in 1964, a brief suborbital test. But the technology spent decades as a curiosity, distrusted for real missions, until Deep Space 1 (1998) flew an ion engine as its main propulsion and proved it could be relied on for years. The pattern is the one this whole book keeps finding: the physics was understood long before the engineering, economics, and nerve caught up.

🔧 Engineering Reality: why xenon? The propellant of choice for both ion and Hall thrusters is xenon, and the reasons are a checklist of practical virtues. It is a noble gas, so it is chemically inert and will not corrode the engine or react with anything. It is easily ionized (its outer electron comes off at only $12.1\ \text{eV}$), which keeps the ionization cost low. It is heavy ($131\ \text{u}$), which — from $v_e = \sqrt{2qV/m}$ — means it reaches a useful $I_{sp}$ at a manageable voltage rather than an extreme one, and gives more thrust per unit power than a light gas would. And it is storable: xenon compresses into a dense, high-pressure supercritical fluid, so hundreds of kilograms fit in a modest tank. Its one flaw is cost — xenon is rare and expensive, which, as we will see, is beginning to push operators toward cheaper gases.

Definition (xenon). Xenon (Xe, atomic number 54, atomic mass $\approx 131.3\ \text{u}$) is a heavy, inert noble gas that is the standard propellant for ion and Hall thrusters: easy to ionize, safe to store as a dense fluid, and heavy enough to give good thrust per watt at practical voltages.

20.3 Hall thrusters: trading a little Isp for a lot more thrust

The gridded ion engine's space-charge limit is frustrating: it is precisely the electric field between the grids, needed to accelerate the ions, that also chokes the current. What if you could accelerate ions without grids — inside a plasma that is already electrically neutral, so there is no space-charge cloud to fight? That is the trick of the Hall thruster, and it is why Hall thrusters, not gridded ion engines, have become the commercial workhorse of the satellite industry.

Definition (Hall thruster). A Hall thruster accelerates ions out of a quasi-neutral plasma using crossed electric and magnetic fields. A radial magnetic field traps electrons so they cannot stream freely to the anode; instead they drift in a circular Hall current and act as a virtual electrode, sustaining a strong axial electric field that accelerates ions out of the channel. Because the plasma is neutral (electrons are present among the ions), there is no space-charge limit, so a Hall thruster produces far more thrust per unit area than a gridded ion engine.

The layout is an annular channel — a short cylindrical trench — with an anode at the bottom (where xenon is injected) and a cathode outside the exit. A magnetic circuit puts a radial magnetic field across the channel near its mouth. Electrons from the cathode are drawn toward the anode, but the magnetic field grabs them: rather than crossing the field lines, they are pinned into an azimuthal drift, circling around the channel in the crossed electric and magnetic fields (an $\mathbf{E}\times\mathbf{B}$ drift). This whirling ring of trapped electrons does two jobs. It bombards and ionizes the incoming xenon, and, because the electrons are held back while the field still pulls, it sets up a steep drop in electric potential right at the channel exit. The heavy xenon ions — unaffected by the magnetic field because they are so much more massive — feel that potential drop and are flung out. Electrons are then fed into the exhaust from the same cathode to neutralize it.

The consequence is a thruster with a very different personality from the gridded ion engine. Because there is no space-charge bottleneck, a Hall thruster pushes far more propellant through a given aperture, so it makes more thrust for the same power. The price is a lower specific impulse: the accelerating potential in a Hall thruster is gentler and less precisely defined than the clean kilovolt of a grid, so the ions leave slower. Typical Hall thrusters run at $I_{sp} \approx 1{,}500$–$2{,}000\ \text{s}$, versus $3{,}000$–$4{,}000\ \text{s}$ for gridded ion engines — still five to ten times a chemical rocket, but half the ion engine's efficiency, in exchange for more push.

Worked Example: comparing two real thrusters at similar power. Consider two flight-proven thrusters:

Thruster	Type	Power	Thrust	$I_{sp}$	$v_e = I_{sp}g_0$	Thrust/power
NSTAR (Dawn)	gridded ion	$2.3\ \text{kW}$	$92\ \text{mN}$	$\sim 3{,}100\ \text{s}$	$\sim 30\ \text{km/s}$	$40\ \text{mN/kW}$
SPT-100	Hall	$1.35\ \text{kW}$	$83\ \text{mN}$	$\sim 1{,}600\ \text{s}$	$\sim 16\ \text{km/s}$	$61\ \text{mN/kW}$

At roughly comparable thrust, the Hall thruster does the job on little more than half the power — its $61\ \text{mN/kW}$ against the ion engine's $40\ \text{mN/kW}$ is a 50% edge in thrust per watt. But look at the specific impulse: the ion engine's $\sim 3{,}100\ \text{s}$ nearly doubles the Hall's $\sim 1{,}600\ \text{s}$, so the ion engine will use about half the propellant to deliver a given delta-v. Let us verify the Hall efficiency from $F = 2\eta P/v_e$: rearranging, $\eta = F v_e / (2P) = (0.083)(15{,}700)/(2 \times 1{,}350) = 0.48$. So the SPT-100 turns about 48% of its electrical power into beam kinetic energy — a believable figure, and the reason a Hall thruster runs hot and needs a good radiator.

This is the essential division of labor in electric propulsion, and it maps neatly onto the trade you first met in Chapter 16 between high thrust and high efficiency. Hall thrusters are the higher-thrust, lower-Isp choice — you reach for one when you want to get somewhere in months rather than years, as in raising a satellite to its operational orbit. Gridded ion engines are the higher-Isp, lower-thrust choice — you reach for one when propellant mass is everything and you have years to spend, as on a deep-space science mission. Neither is "better"; they sit at different points on the same power–thrust–Isp curve.

⚠️ Common Misconception: "A Hall thruster is just a worse ion engine." They are cousins — both ionize xenon and expel ions electrically — but a Hall thruster is not a degraded ion engine; it is a deliberately different point on the trade. By giving up some specific impulse it escapes the space-charge limit and wins a large increase in thrust density and simplicity (no delicate grids to erode or misalign). For a comsat that must raise its orbit before it can earn revenue, the Hall thruster's faster transfer is worth far more than the ion engine's propellant savings. The "worse" number ($I_{sp}$) is the right number to sacrifice for that mission. Choosing which figure of merit to optimize is the engineering.

🔄 Check Your Understanding 1. What physical mechanism lets a Hall thruster avoid the space-charge limit that caps a gridded ion engine's current? 2. You have a fixed 3 kW of power. You want the most thrust to raise an orbit quickly. Ion or Hall — and why?

Answers

1. The Hall thruster accelerates ions inside a quasi-neutral plasma — the electrons are mixed in with the ions, so there is no net space charge to build up a repulsive field and choke the current. The gridded ion engine, by contrast, accelerates a beam of bare positive ions across a vacuum gap, where their mutual repulsion (space charge) limits the current. 2. A Hall thruster: at fixed power it delivers more thrust (lower $v_e$ means higher $F = 2\eta P/v_e$), so it raises the orbit faster. You pay in propellant (lower $I_{sp}$), but for a time-critical orbit raise that is the right trade.

20.4 The rest of the family: PPT, MPD, and VASIMR

Ion and Hall thrusters dominate flown hardware, but they are two members of a larger family. It helps to sort electric thrusters by how they accelerate the propellant, because the physics of the accelerator sets the performance.

• Electrostatic (ion, Hall): accelerate charged ions with an electric field. This is where the workhorses live.
• Electromagnetic (MPD, PPT): accelerate a current-carrying plasma with the $\mathbf{J}\times\mathbf{B}$ Lorentz force — the plasma's own current, crossed with a magnetic field, pushes it out.
• Electrothermal (resistojet, arcjet): simply use electricity to heat a propellant, then expand it through an ordinary nozzle. These are the modest, reliable low-Isp end of the family — a resistojet is little more than an electric kettle with a nozzle — and we will not dwell on them.

Three of the more ambitious concepts are worth meeting by name, because you will hear them invoked whenever the conversation turns to the future of in-space propulsion.

Pulsed plasma thrusters (PPTs) are the simplest electric rockets that exist, and among the oldest flown — a Soviet PPT reached space aboard Zond-2 in 1964. A PPT uses a solid propellant, usually a bar of Teflon (PTFE). A capacitor discharges an arc across the face of the bar; the arc ablates and ionizes a puff of Teflon, and the same discharge current, crossed with its self-induced magnetic field, blows the little plasma cloud out electromagnetically. Each pulse delivers a minuscule, precisely repeatable bit of impulse — micronewton-seconds — which makes PPTs excellent for the delicate attitude control and precise positioning of small satellites, at the cost of very low average thrust and power.

Magnetoplasmadynamic (MPD) thrusters are the muscle of the electromagnetic family. A very high current (thousands of amps) arcs through a plasma; the current's own magnetic field crosses the current to produce a strong $\mathbf{J}\times\mathbf{B}$ force that accelerates the plasma to $20$–$50\ \text{km/s}$. Unlike ion and Hall thrusters, an MPD thruster is not space-charge limited and can, in principle, produce thrust measured in whole newtons — orders of magnitude more than a Hall thruster. The catch is written in its power appetite: useful MPD operation wants hundreds of kilowatts to megawatts, far beyond what today's solar arrays can supply. MPD is a technology waiting for a power source, which is exactly why it recurs in discussions of nuclear-electric propulsion in Chapter 21.

VASIMR — the Variable Specific Impulse Magnetoplasma Rocket — is the most talked-about of the advanced concepts. It uses radio waves to do everything: one antenna ionizes a gas (usually argon) into a plasma, a second antenna heats that plasma to enormous temperatures, and a magnetic nozzle — shaped magnetic field lines rather than a metal bell — lets the hot plasma expand and push. Because nothing touches the plasma (no grids, no electrodes to erode), VASIMR promises long life, and because you can adjust how much you heat the plasma, it can trade thrust for specific impulse in flight — high thrust and low $I_{sp}$ when you want to move quickly, low thrust and high $I_{sp}$ when you want efficiency. That variability is genuinely appealing.

🔧 Engineering Reality: separate the physics from the press release. MPD and VASIMR are real, demonstrated in the laboratory, and grounded in sound plasma physics — but both are throttled by the same wall: power. A VASIMR unit has been run at $200\ \text{kW}$ on a test stand, but a spacecraft that could supply $200\ \text{kW}$ continuously does not yet exist, because the solar arrays or reactor to make it would themselves weigh tonnes. Neither MPD nor VASIMR has flown as a mission's main propulsion. This is not a knock on the concepts; it is a reminder that in electric propulsion the thruster is rarely the hard part — the power plant is. When you read that a technology will "get us to Mars in weeks," check what power source it assumes, and whether that source exists. The honest status of high-power electric propulsion is: promising physics, waiting on Chapter 25's power systems and Chapter 21's reactors to catch up.

20.5 The thrust–efficiency tradeoff, quantified

We have met the central relationship twice now; here we make it earn its keep. Everything an electric thruster can and cannot do is contained in

$$ F = \frac{2\,\eta\,P}{v_e}, \qquad \dot m = \frac{F}{v_e}, \qquad P_{\text{jet}} = \tfrac{1}{2}\dot m\,v_e^2 = \eta P. $$

Let us define the figure of merit that names this chapter's key term.

Definition (power-to-thrust ratio). The power-to-thrust ratio is the electrical power a thruster must draw per unit of thrust it produces, $P/F$. From $F = 2\eta P/v_e$, it is $$\frac{P}{F} = \frac{v_e}{2\eta}.$$ It rises directly with exhaust velocity: the higher your specific impulse, the more power you must spend for each newton of thrust. (Its reciprocal, thrust-to-power $F/P$ in mN/kW, is the number satellite engineers usually quote.)

Put NSTAR's numbers in: $P/F = v_e/(2\eta) = 30{,}400/(2 \times 0.61) \approx 25{,}000\ \text{W/N}$, or 25 kilowatts per newton. Sit with that figure. To produce a single newton of thrust — the weight of a small apple — an ion engine of this class needs $25\ \text{kW}$, about the power of ten household kitchens. To produce the $10\ \text{N}$ you might want for even a leisurely maneuver, you would need a quarter of a megawatt. This is the quantitative face of "low thrust," and it leads directly to the question every newcomer asks.

⚠️ Common Misconception: "If electric propulsion is so efficient, why don't we launch rockets with it?" Because thrust scales with power, and the power you would need is physically impossible to carry. Let us prove it. To lift a modest $1{,}200\ \text{kg}$ spacecraft off Earth, thrust must at least equal weight: $F > mg = 1{,}200 \times 9.81 = 11{,}800\ \text{N}$. One NSTAR-class ion engine makes about $0.092\ \text{N}$, so you would need $11{,}800/0.092 \approx 128{,}000$ engines. At $2.3\ \text{kW}$ each, that is $128{,}000 \times 2.3\ \text{kW} \approx 294\ \text{MW}$ — the output of a fair-sized power station. The solar array to generate $294\ \text{MW}$ in space, at roughly $100\ \text{W/kg}$, would mass about $3{,}000\ \text{tonnes}$ — two thousand times the spacecraft it is trying to lift. Electric propulsion cannot launch not because engineers have not tried hard enough, but because a thrust-to-weight ratio far below one is baked into the physics. It works only in space, where you never need to fight gravity by brute hovering — you fight it patiently, with orbital mechanics.

Here is the deeper reason the tiny thrust is unavoidable, framed with the thrust equation from Chapter 16. Thrust is $F = \dot m\, v_e$. Chemical and electric rockets reach the same thrust from opposite directions. A Merlin engine flows propellant at roughly $\dot m \approx 300\ \text{kg/s}$ at $v_e \approx 3{,}000\ \text{m/s}$, giving $F \approx 900{,}000\ \text{N}$. An ion engine flows xenon at $\dot m \approx 3\ \text{mg/s} = 3\times10^{-6}\ \text{kg/s}$ at $v_e \approx 30{,}000\ \text{m/s}$, giving $F \approx 0.09\ \text{N}$. The exhaust velocities differ by ten; the mass flows differ by a hundred million. That is the whole story in two numbers: the chemical rocket throws a river of propellant at moderate speed; the ion engine throws a thread of it at tremendous speed. Same physics, same equation, opposite regimes.

Worked Example: thrust and flow of an ion engine from power and Isp. Predict the thrust of a thruster drawing $P = 2.3\ \text{kW}$ at $I_{sp} = 3{,}100\ \text{s}$ and efficiency $\eta = 0.61$, and check it against the real NSTAR (92 mN).

First the exhaust velocity: $v_e = I_{sp}\,g_0 = 3{,}100 \times 9.81 = 30{,}411\ \text{m/s}$. Then the thrust: $$F = \frac{2\eta P}{v_e} = \frac{2 \times 0.61 \times 2{,}300}{30{,}411} = \frac{2{,}806}{30{,}411} = 0.0923\ \text{N} \approx 92\ \text{mN}. \quad\checkmark$$ The propellant mass flow: $$\dot m = \frac{F}{v_e} = \frac{0.0923}{30{,}411} = 3.0\times10^{-6}\ \text{kg/s} = 3.0\ \text{mg/s}.$$ Three milligrams of xenon per second, thrown at $30\ \text{km/s}$, to push $92\ \text{mN}$. Sanity check on the acceleration: on Dawn's $\sim 1{,}200\ \text{kg}$, that thrust gives $a = F/m = 0.092/1200 = 7.7\times10^{-5}\ \text{m/s}^2$, about eight *millionths* of a $g$. The Dawn team liked to say the thrust felt like the weight of a single sheet of paper resting on your hand. And yet, run for years, it delivered more delta-v than any spacecraft before it.

Now, the trade-off has a subtler edge that matters for real design. Because power hardware has mass — a solar array massive enough to feed the thruster is not free — you cannot simply crank $v_e$ to the sky. Push the specific impulse higher and the propellant mass shrinks (good), but the power-to-thrust ratio rises, so either you add power-generation mass (bad) or you accept even less thrust and a longer mission (also bad, if boiloff, radiation dose, or a deadline is pressing). Somewhere between "too low an $I_{sp}$, wasting propellant" and "too high an $I_{sp}$, drowning in power-system mass and time" lies an optimum. For a great many real missions that optimum lands in the $1{,}500$–$4{,}000\ \text{s}$ range — which is, not by coincidence, exactly where Hall and ion thrusters were built to operate. The hardware was shaped by the missions.

🔗 Connection: This is why the power source is inseparable from the propulsion choice. The array or reactor that feeds an electric thruster is sized, degraded by radiation, and eclipsed by the planet exactly as Chapter 25 will describe; for a deep-space mission far from the Sun, where sunlight is faint, you may need the radioisotope or reactor power of Chapter 25 and the nuclear-electric ideas of Chapter 21 before an electric thruster is even an option. In electric propulsion, "what engine?" and "what power source?" are one question.

🔄 Check Your Understanding 1. An engineer proposes doubling a thruster's $I_{sp}$ "to save propellant," keeping the same power. Name one cost of doing so. 2. Why is it the mass flow rate, not the exhaust velocity, that makes electric thrust so small?

Answers

1. Thrust halves (from $F = 2\eta P/v_e$), so every maneuver takes twice as long — and if the mission has a deadline, or the spacecraft accrues radiation dose or propellant boiloff over time, that extra time can cost more than the propellant it saves. (Higher $v_e$ also demands higher voltage and can lower efficiency.) 2. Because thrust is $F = \dot m v_e$, and while an ion engine's $v_e$ is large (~10× chemical), its $\dot m$ is tiny (milligrams per second, ~$10^8$ times smaller than a chemical engine's). The small mass flow, set by the space-charge/power limit, dominates — a huge exhaust velocity cannot rescue a thread-thin mass flow.

20.6 Spiral transfers and where electric propulsion wins

A chemical rocket's burn is so brief compared with an orbit that we model it as an instantaneous kick — an impulsive maneuver, as in the Hohmann transfer of Chapter 10. An electric thruster cannot pretend to be impulsive. Its burn lasts weeks or months, spanning hundreds or thousands of orbital revolutions, so the thrust must be treated as it truly is: a continuous, gentle push acting over a huge arc of the trajectory. This changes the shape of the path entirely.

Definition (low-thrust trajectory). A low-thrust trajectory is one flown with thrust so small compared with the local gravitational force that the burn cannot be approximated as an instantaneous velocity change. The engine thrusts continuously over a large fraction of the trajectory, and the path must be found by integrating the motion under thrust plus gravity rather than by summing impulses.

Definition (spiral transfer). A spiral transfer is the low-thrust way to change orbits: the spacecraft thrusts continuously (typically along its velocity vector, to add or remove energy most efficiently) and slowly winds outward — or inward — through many nearly circular revolutions, tracing a spiral from the starting orbit to the target orbit.

The natural question is what such a spiral costs in delta-v, so we can compare it against the impulsive Hohmann transfer we already know how to price. For the clean case — a continuous, tangential-thrust spiral between two coplanar circular orbits — there is a beautifully simple answer (it is Edelbaum's result): the delta-v is just the difference of the two circular orbital speeds,

$$ \Delta v_{\text{spiral}} = \lvert v_1 - v_2 \rvert, \qquad v = \sqrt{\frac{\mu}{r}}, $$

using the circular velocity from Chapter 6. Let us compare the two transfer strategies on the most important commercial route in spaceflight: from low Earth orbit up to geostationary orbit (Chapter 9).

Worked Example: spiral vs. Hohmann, LEO to GEO. Take a $400\ \text{km}$ LEO ($r_1 = 6{,}778\ \text{km}$) and GEO ($r_2 = 42{,}164\ \text{km}$), with Earth's $\mu = 3.986\times10^{5}\ \text{km}^3/\text{s}^2$.

Circular speeds: $$v_1 = \sqrt{398{,}600/6{,}778} = 7.669\ \text{km/s}, \qquad v_2 = \sqrt{398{,}600/42{,}164} = 3.075\ \text{km/s}.$$

Spiral delta-v (continuous low thrust): $$\Delta v_{\text{spiral}} = v_1 - v_2 = 7.669 - 3.075 = 4.59\ \text{km/s}.$$

Hohmann delta-v (two impulsive burns, via the transfer ellipse with $a_t = (r_1+r_2)/2 = 24{,}471\ \text{km}$): the perigee burn is $\sqrt{\mu(2/r_1 - 1/a_t)} - v_1 = 10.067 - 7.669 = 2.40\ \text{km/s}$, and the apogee burn is $v_2 - \sqrt{\mu(2/r_2 - 1/a_t)} = 3.075 - 1.618 = 1.46\ \text{km/s}$. Adding the (unrounded) burns, $$\Delta v_{\text{Hohmann}} = 3.85\ \text{km/s}.$$

The spiral costs more delta-v — $4.59$ versus $3.85\ \text{km/s}$, about 19% more. That is the price of thrusting continuously, much of it at angles and altitudes where an impulsive burn would not waste effort.

So the low-thrust spiral is the less efficient route in pure delta-v. Why on Earth would anyone use it? Because delta-v is not the bill you pay — propellant mass is, and the electric thruster's tenfold specific impulse crushes the propellant even while paying the delta-v surcharge. Let us finish the comparison with the rocket equation.

Worked Example: which route needs less xenon (or kerosene)? A $1{,}200\ \text{kg}$ (final-mass) satellite must climb from LEO to GEO.

Chemical, Hohmann ($\Delta v = 3.85\ \text{km/s}$, $I_{sp} = 320\ \text{s}$, $v_e = 3{,}138\ \text{m/s}$): $$m_p = m_f\left(e^{\Delta v/v_e} - 1\right) = 1{,}200\left(e^{3850/3138} - 1\right) = 1{,}200\,(3.42 - 1) = 2{,}900\ \text{kg}.$$

Electric, spiral ($\Delta v = 4.59\ \text{km/s}$, $I_{sp} = 1{,}800\ \text{s}$ Hall, $v_e = 17{,}652\ \text{m/s}$): $$m_p = 1{,}200\left(e^{4590/17652} - 1\right) = 1{,}200\,(1.297 - 1) = 356\ \text{kg}.$$

The chemical route burns $2{,}900\ \text{kg}$; the electric route, despite needing more delta-v, burns only $356\ \text{kg}$ — an eight-fold propellant saving. On a satellite launched at, say, $3{,}500\ \text{kg}$, that is roughly $2{,}500\ \text{kg}$ you no longer have to lift — the difference between a big, expensive launch and a small one, or between one satellite and two sharing a ride.

There is, inevitably, a cost, and by now you can guess its name: time. That $356\ \text{kg}$ of xenon, expelled at the milligram-per-second rate a Hall thruster manages, does not leave quickly.

Worked Example: how long is the spiral? A $4.5\ \text{kW}$ Hall thruster at $I_{sp} = 1{,}800\ \text{s}$ ($v_e = 17{,}652\ \text{m/s}$) and $\eta = 0.55$ produces $$F = \frac{2\eta P}{v_e} = \frac{2(0.55)(4{,}500)}{17{,}652} = 0.280\ \text{N} = 280\ \text{mN}.$$ Its mass flow is $\dot m = F/v_e = 0.280/17{,}652 = 1.59\times10^{-5}\ \text{kg/s}$, so expelling $356\ \text{kg}$ takes $$t = \frac{m_p}{\dot m} = \frac{356}{1.59\times10^{-5}} = 2.24\times10^{7}\ \text{s} \approx 260\ \text{days}.$$ Roughly eight months of near-continuous thrusting to climb to GEO — versus a chemical transfer of a few hours. That is the deal electric propulsion offers, in one sentence: give me months instead of hours, and I will get you there on a fraction of the propellant.

Whether that deal is worth taking is the essence of choosing a propulsion system, and the answer depends entirely on the mission. Three real programs show the spread.

Dawn (NASA, 2007–2018) is the mission electric propulsion made possible outright. To orbit the giant asteroid Vesta, study it, then leave and orbit the dwarf planet Ceres, Dawn needed on the order of $11\ \text{km/s}$ of delta-v from its own engines — a budget so large that a chemical spacecraft to do it would have been mostly propellant and far too heavy to launch affordably. Dawn carried three NSTAR ion engines and $425\ \text{kg}$ of xenon; over years of patient thrusting it accumulated about $11.5\ \text{km/s}$, more than any spacecraft before it, and became the first ever to orbit two separate extraterrestrial bodies. (A rocket-equation check: $v_e \ln(m_0/m_f) = 30{,}400 \times \ln(1{,}218/793) \approx 13\ \text{km/s}$ if every gram of xenon were spent at full $I_{sp}$ — comfortably above the $11.5$ actually realized, the balance lost to throttling and reserves.) No chemical spacecraft of Dawn's launch mass could have visited even one of its targets, let alone both.

Starlink (SpaceX) shows electric propulsion at industrial scale. Every one of the thousands of Starlink satellites carries a Hall thruster that raises it from its drop-off altitude to operational orbit, holds its slot against drag and perturbations (Chapter 12), dodges conjunctions, and finally deorbits it responsibly at end of life. The economics are ruthless and instructive: the first Starlink generation used krypton rather than xenon — a lighter, cheaper noble gas that gives slightly higher $I_{sp}$ at somewhat lower thrust and efficiency — and later versions moved to cheaper argon still, trading thruster efficiency for propellant cost because, at a scale of ten thousand satellites, the price of the gas dominates. This is the propellant-selection logic of the whole chapter, playing out as a business decision (a story we return to with constellations in Chapter 33).

BepiColombo (ESA/JAXA) aims the technology at one of the solar system's hardest destinations: Mercury. Reaching a low orbit at Mercury means shedding an enormous amount of the speed you inherit falling toward the Sun — a delta-v so steep that BepiColombo combines solar-electric ion propulsion with a long series of planetary gravity assists (Chapter 11). Its Mercury Transfer Module carries four gridded ion thrusters and thrusts for months at a time across a multi-year cruise, braking the spacecraft against the Sun's pull with a patience no chemical stage could sustain. It is the clearest demonstration that for the highest-delta-v missions, low thrust applied for a very long time is not a handicap — it is the only affordable way there.

📜 From History: The first operational use of electric propulsion was humbler and older than any of these: north–south station-keeping on geostationary communications satellites, which began adopting electric thrusters in the 1980s–1990s. A GEO satellite must spend roughly $50\ \text{m/s}$ of delta-v every year fighting the Sun and Moon's tug on its orbit; over a 15-year life that is $\sim 0.75\ \text{km/s}$, and doing it with high-$I_{sp}$ electric thrusters instead of chemical ones saved enough propellant mass to add years of life or payload. That unglamorous application — nudging a satellite a few meters per second per week — quietly proved the technology and paid for its development, long before it flew to the asteroids.

Mission Design Checkpoint: should your mission fly electric?

This chapter adds a decision to your Mission Design Review, and an electric-propulsion increment to the astrotools propulsion.py module you began building in Chapters 16–19.

The design decision. Open your MDR and ask, honestly, whether electric propulsion fits your mission. The rule of thumb from this chapter: electric propulsion wins when delta-v is large and time is cheap; it loses when you need high thrust or a fast maneuver. Match it to your track:

• Track A — GEO comsat: a strong fit. Use a Hall thruster for all-electric orbit-raising from your launch drop-off to GEO (accepting a several-month spiral to save thousands of kilograms of launch mass) and for the years of north–south station-keeping. Note in your MDR the propellant saved versus a chemical apogee kick — and the orbit-raising time you are accepting.
• Track D — asteroid rendezvous: a strong fit, for exactly Dawn's reasons. The delta-v to match a small body's orbit and then maneuver around it is large, and you have years of cruise to spend; ion propulsion may be the only way to close the mass budget. Record a gridded-ion option.
• Track C — Mars orbiter: a possible fit for the cruise (solar-electric transfer), trading a longer flight for a smaller launch, though many Mars orbiters still use chemical capture for its speed.
• Track B — lunar lander: electric propulsion cannot perform the landing (you need high thrust to hover and touch down — that is chemical territory), but it can deliver the vehicle to lunar orbit by spiraling out from Earth, exactly as ESA's SMART-1 reached the Moon on a Hall thruster over about thirteen months. Note that split: electric cruise, chemical descent.

Write one paragraph in your MDR stating your choice, your chosen thruster type and rough $I_{sp}$, and — if electric — the transfer time you are accepting. You will size the propellant against your delta-v budget in Chapter 29.

The code. Add the electric-propulsion helpers to astrotools/propulsion.py:

import math

G0 = 9.80665  # standard gravity, m/s^2 (unit-conversion constant)

def ep_thrust(power_w, isp_s, efficiency):
    """Thrust (N) of an electric thruster: F = 2*eta*P / v_e, with v_e = Isp*g0."""
    ve = isp_s * G0
    return 2 * efficiency * power_w / ve

def ep_mass_flow(thrust_n, isp_s):
    """Propellant mass flow (kg/s): mdot = F / v_e."""
    return thrust_n / (isp_s * G0)

def ep_burn_time(prop_kg, thrust_n, isp_s):
    """Continuous-thrust time (s) to expend prop_kg of propellant: t = m_p / mdot."""
    return prop_kg / ep_mass_flow(thrust_n, isp_s)

if __name__ == "__main__":
    # Track A comsat: all-electric LEO->GEO raise with a 4.5 kW Hall thruster.
    power, isp, eta = 4500, 1800, 0.55
    thrust = ep_thrust(power, isp, eta)
    print("thrust:", round(thrust * 1000, 1), "mN")
    # Xenon to supply the 4.59 km/s spiral for a 1200 kg satellite:
    ve = isp * G0
    xenon = 1200 * (math.exp(4590 / ve) - 1)
    print("xenon needed:", round(xenon), "kg")
    print("burn time:", round(ep_burn_time(xenon, thrust, isp) / 86400), "days")
    # Expected output:
    # thrust: 280.4 mN
    # xenon needed: 356 kg
    # burn time: 260 days

The output ties the whole chapter together: a $4.5\ \text{kW}$ Hall thruster makes $280\ \text{mN}$, sips $356\ \text{kg}$ of xenon to spiral a comsat to GEO, and takes about $260$ days to do it. Point these functions at your mission's numbers and they will tell you the thrust you can expect, the propellant you must carry, and the months you must budget.

Summary

Electric propulsion is a direct assault on the tyranny of the rocket equation: by decoupling the energy source from the reaction mass, it lifts the chemical ceiling on exhaust velocity, at the price of tiny thrust. It is also, quietly, one of the technologies reshaping what is affordable in space — halving the launch mass of a communications satellite, making ten-thousand-satellite constellations economically viable — a transformation that, like reusability, we are only at the beginning of. Carry these forward:

Numbers worth remembering: electric $I_{sp}$ is $1{,}500$–$4{,}000+\ \text{s}$ (vs. chemical $250$–$460$); thrust is tens to hundreds of milli-newtons; a LEO→GEO spiral is $\sim 4.6\ \text{km/s}$ over months; xenon is the standard propellant (krypton/argon when cost rules).

Spaced Review

Retrieval strengthens memory. Answer from memory before checking, then look back at the cited chapter.

1. (Ch. 16) State the thrust equation $F = \dot m\, v_e$ in words, and use it to explain why an ion engine ($v_e \approx 30\ \text{km/s}$) produces far less thrust than a Merlin ($v_e \approx 3\ \text{km/s}$), even though its exhaust is ten times faster.
2. (Ch. 16) Chapter 16 framed a trade between high thrust and high efficiency. Which end of that trade does electric propulsion sit at, and what physical quantity are you sacrificing to get there?
3. (Ch. 3) An electric thruster has $I_{sp} = 2{,}500\ \text{s}$. What is its exhaust velocity in km/s? Roughly how many times a good chemical engine ($v_e \approx 3.3\ \text{km/s}$) is that?
4. (Ch. 3) A spacecraft must do $\Delta v = 4\ \text{km/s}$. Compare the mass ratio needed with a chemical engine ($v_e = 3.3\ \text{km/s}$) and an electric one ($v_e = 25\ \text{km/s}$). Why does the electric engine barely change the vehicle's mass?

Answers

1. Thrust equals the mass flow rate times the exhaust velocity — the rate at which the engine throws momentum backward. The ion engine's $v_e$ is ~10× higher, but its mass flow $\dot m$ is ~$10^8$ times smaller (milligrams vs. hundreds of kilograms per second), and the product $\dot m v_e$ is therefore vastly smaller. Thrust is dominated by how much mass you throw, not just how fast. 2. Electric propulsion sits at the high-efficiency (high-$I_{sp}$) end; it sacrifices thrust (and thus the ability to maneuver quickly or launch) to get there. 3. $v_e = 2{,}500 \times 9.81 \approx 24{,}500\ \text{m/s} = 24.5\ \text{km/s}$ — about $7$–$8$ times a good chemical engine. 4. Chemical: $m_0/m_f = e^{4000/3300} = e^{1.21} \approx 3.36$ (so ~70% propellant). Electric: $m_0/m_f = e^{4000/25000} = e^{0.16} \approx 1.17$ (so ~15% propellant). Because $v_e$ sits in the denominator of the exponent, the electric engine's huge $v_e$ makes the exponent small, so the mass ratio barely exceeds 1 — almost no propellant is needed.

What's Next

Electric propulsion broke the chemical ceiling on exhaust velocity by moving the energy source off the propellant and onto the power system — and then ran headlong into a new ceiling: power. The thrust of an ion engine, an MPD, a VASIMR is throttled not by the thruster but by how many watts the spacecraft can generate and carry. Solar arrays run out of sunlight in the outer solar system and cannot easily reach the hundreds of kilowatts that MPD and VASIMR crave. So the natural next question is: where else can we get energy, and can we get it in the enormous quantities that would make high-thrust and high-Isp possible at once? In Chapter 21 we turn to nuclear and advanced propulsion — nuclear-thermal rockets that heat hydrogen in a reactor, nuclear-electric systems that feed the very thrusters of this chapter with reactor power, and the further frontier of solar sails, beamed power, and antimatter — separating, as always, the physics of the possible from the engineering of the probable.