Chapter 18: Combustion and Propellants

"One test result is worth one thousand expert opinions." — Wernher von Braun

Overview

In Chapter 16 we made specific impulse rigorous, and in Chapter 17 we took the engine apart — injector, chamber, throat, turbopumps, and the cycles that feed them. But we treated the fire itself as a black box: propellant goes in, hot gas comes out at some exhaust velocity $v_e$, and the rocket equation does the rest. This chapter opens that black box. What actually happens in the few milliseconds a propellant molecule spends in a combustion chamber, and why does that chemistry set the ceiling on everything a chemical rocket can do?

The answer comes down to two numbers, and by the end of this chapter you will see every rocket engine in terms of them. The chemistry of the fire sets a flame temperature — how hot the gas gets — and an exhaust molecular weight — how heavy the gas molecules are. Exhaust velocity depends on the ratio of these two: hotter is better, lighter is better, and $v_e$ scales like the square root of temperature divided by molecular weight. That single fact explains the whole periodic table of rocket fuels: why hydrogen is the efficiency champion even though it does not burn the hottest, why kerosene and methane trade efficiency for density, and why no chemical rocket will ever reach the exhaust velocities of an ion engine. The chemistry is the reason the tyranny of the rocket equation (Chapter 3) bites as hard as it does.

Then we confront the field's most notorious hard problem. A rocket combustion chamber is the most power-dense sustained energy-release device humans build — a Raptor chamber liberates energy at a rate comparable to a large nuclear power station, in a volume smaller than a beer keg. Get the chemistry, the plumbing, and the acoustics even slightly out of tune and the fire can begin to sing — pressure oscillations that feed on themselves and destroy the engine in milliseconds. This is combustion instability, and it nearly sank the Apollo program. It is the sharpest expression in this book of Theme 2: space is an unforgiving environment, and everything must work.

In this chapter, you will learn to:

• Read a combustion reaction and estimate the energy it releases, per kilogram of propellant.
• Explain the two knobs — flame temperature and exhaust molecular weight — that set exhaust velocity, and use them to predict which propellant wins.
• Say precisely why engines burn fuel-rich, and why hydrogen's light exhaust beats kerosene's hotter fire.
• Compare the four propellant families that fly today, with their real mixture ratios and trade-offs.
• Explain cryogenics, storability, and why SpaceX chose methane for Starship.
• Describe combustion instability, the Rayleigh criterion behind it, and how engineers tame it.

Learning Paths

🚀 Space Enthusiast: Read 18.1 for the idea, then spend your time on 18.3 (the propellant families — the part you can bring to a launch watch party) and 18.5 (combustion instability, the best story in propulsion). You can skim the algebra of 18.2 and still get its punchline: hotter and lighter is better.

📐 Engineering Student: Read everything. Section 18.2 is the analytical heart — the $\sqrt{T_c/\mathcal{M}}$ scaling underlies all of Chapters 18–19, and the fuel-rich optimization in 18.3 is a favorite exam topic. Do the ⭐⭐/⭐⭐⭐ exercises.

🎮 KSP Player: Your game hides this chapter inside each engine's Isp stat. Focus on 18.3 and 18.4 to understand why the LV-909 (hydrogen-like, high Isp, low thrust) and the Mainsail (kerosene-like, high thrust) feel so different, and why real vehicles mix them.

🛰️ Industry Prep: Sections 18.3, 18.4, and the Mission Design Checkpoint are propellant selection — a real trade study you will run for your mission. Green propellants (18.6) are where operations cost and regulation are moving fastest.

18.1 Combustion chemistry and the energy released

Everything a chemical rocket does begins with a fire — a very fast, very hot, very well-behaved one. Before we can talk about temperature or exhaust speed, we have to be clear about what combustion actually is and where its energy comes from.

Definition (combustion). Combustion is a rapid, self-sustaining, exothermic oxidation–reduction reaction between a fuel and an oxidizer that releases stored chemical energy as heat, producing hot gaseous products. In a rocket, that heat is the entire point: it is the thermal energy a nozzle will convert into the directed kinetic energy of the exhaust.

The word oxidation is a clue to where the energy lives. Chemical energy is stored in the bonds between atoms, and different bonds store different amounts. A combustion reaction breaks the relatively weak bonds of the reactants (the H–H bond in hydrogen, the O=O bond in oxygen) and forms the much stronger, lower-energy bonds of the products (the O–H bonds in water). Nature always "prefers" the lower-energy configuration, and the energy difference has to go somewhere — it comes out as the kinetic energy of the product molecules, which is to say, as heat. A rocket is a machine for cashing in that bond-energy difference and pointing the proceeds at the ground.

Bookkeeping the energy: enthalpy of formation. Chemists track this with the enthalpy of formation $\Delta H_f$, the energy absorbed or released when a compound is built from its elements in their standard states. Elements in their natural form (H$_2$, O$_2$) have $\Delta H_f = 0$ by definition; stable compounds like water have large negative $\Delta H_f$, meaning energy was released to form them. The heat released by a reaction is just the difference between where you end and where you started:

$$ \Delta H_{\text{rxn}} = \sum \Delta H_f(\text{products}) - \sum \Delta H_f(\text{reactants}). $$

Let us make it concrete with the simplest and most important rocket reaction, hydrogen burning in oxygen:

$$ 2\,\text{H}_2 + \text{O}_2 \;\rightarrow\; 2\,\text{H}_2\text{O}. $$

Worked Example: how much energy does hydrogen combustion release? The enthalpy of formation of gaseous water (the relevant phase — the exhaust leaves as superheated vapor) is $\Delta H_f \approx -241.8\ \text{kJ/mol}$. The reactants are elements, so their $\Delta H_f = 0$. Forming two moles of water therefore releases $$\Delta H_{\text{rxn}} = 2 \times (-241.8) - 0 = -483.6\ \text{kJ},$$ the minus sign meaning energy leaves the reaction. Now weigh the propellant that produced it: two moles of H$_2$ ($2 \times 2.016 = 4.03\ \text{g}$) plus one mole of O$_2$ ($32.00\ \text{g}$) is $36.03\ \text{g}$, or $0.03603\ \text{kg}$. The energy released per kilogram of propellant burned is $$\frac{483.6\ \text{kJ}}{0.03603\ \text{kg}} \approx 13.4\ \text{MJ/kg}.$$ Sanity check: TNT releases about $4.6\ \text{MJ/kg}$, so a kilogram of stoichiometric hydrogen–oxygen carries roughly three times the energy of a kilogram of TNT. That should feel right: a rocket is a bomb we have persuaded to burn steadily out one end instead of all at once.

Here is the first deep lesson of the chapter, and it is a humbling one. Run the same calculation for the other propellants and you get numbers that are startlingly close together — not because the fuels are similar, but because we must carry the oxidizer too, and oxidizer is heavy dead weight in the energy accounting:

Hydrogen looks like a miracle fuel per kilogram of fuel — and it is — but once you include the eight kilograms of oxygen needed to burn each kilogram of it, the advantage per kilogram of propellant shrinks to a modest ~35%. Every practical chemical propellant releases something like $10\text{–}13\ \text{MJ/kg}$ of mixture. That narrow band is chemistry's hard ceiling, and it is exactly why the specific impulses you met in Chapter 3 were all bunched within a factor of two. You cannot escape the rocket equation by finding a magically more energetic fuel; there isn't one. (These heats of combustion are standard textbook values, Tier 2 — good to about two significant figures.)

Definition (stoichiometry). Stoichiometry is the quantitative accounting of a reaction, fixed by the conservation of atoms: you balance the equation so that every atom on the left reappears on the right. The stoichiometric proportion of oxidizer to fuel is the exact ratio that consumes both completely, leaving no excess of either. For $2\,\text{H}_2 + \text{O}_2 \rightarrow 2\,\text{H}_2\text{O}$, burning $4.03\ \text{g}$ of hydrogen needs $32.00\ \text{g}$ of oxygen — a stoichiometric mass ratio of about $8$ to $1$.

That "8 to 1" is our first glimpse of the mixture ratio, the single most important dial an engine designer turns, which we develop fully in 18.3. But to understand why engines almost never run at the stoichiometric ratio, we first need to see what temperature the fire reaches — and why "as hot as possible" turns out to be the wrong goal.

🔄 Check Your Understanding 1. Where, physically, does the energy released in combustion come from? 2. Hydrogen's heat of combustion per kilogram of fuel is nearly three times methane's, yet per kilogram of propellant mixture the two are within ~35%. What accounts for the shrinkage?

Answers

1. From the difference in chemical bond energy between reactants and products: combustion replaces weaker reactant bonds (H–H, O=O) with stronger product bonds (O–H), and the surplus energy emerges as the heat (random kinetic energy) of the product gas. 2. Because a rocket must carry its own oxidizer. Hydrogen needs ~8 kg of oxygen per kg of fuel, so most of the mass you accelerate is oxygen, which dilutes hydrogen's spectacular per-kilogram-of-fuel energy down to a per-kilogram-of-mixture value close to the others.

18.2 Adiabatic flame temperature and exhaust molecular weight

We now have the energy. Where does it go? In an idealized combustion chamber, no heat escapes to the walls (the process is fast and the chamber is, to first order, insulated by its own boundary layer), so all the released energy goes into heating the product gas. The temperature it reaches is the chamber's defining number.

Definition (adiabatic flame temperature). The adiabatic flame temperature $T_c$ is the temperature the combustion products reach when all the released chemical energy goes into heating them, with none lost to the surroundings ("adiabatic" means no heat exchange). It is the maximum temperature the reaction can produce, and it is the $T_c$ that appears in every performance formula from here to Chapter 19.

A first, naive estimate — and a surprise. If the released energy $\Delta H_{\text{rxn}}$ simply heats the products, then energy balance gives $\Delta H_{\text{rxn}} = n\,\bar{c}_p\,(T_c - T_0)$, where $n$ is the moles of product, $\bar{c}_p$ is their average molar heat capacity, and $T_0$ is the starting temperature. Solve for the temperature rise:

$$ T_c \approx T_0 + \frac{\Delta H_{\text{rxn}}}{n\,\bar{c}_p}. $$

Worked Example: the naive flame temperature of hydrogen–oxygen. From 18.1, burning to two moles of water releases $483.6\ \text{kJ}$. Steam at high temperature has a molar heat capacity of roughly $\bar{c}_p \approx 55\ \text{J/(mol·K)}$. With $n = 2\ \text{mol}$ of product and $T_0 = 298\ \text{K}$: $$T_c \approx 298 + \frac{483{,}600\ \text{J}}{2 \times 55\ \text{J/K}} = 298 + 4{,}396 \approx 4{,}700\ \text{K}.$$ But the measured chamber temperature of a hydrogen–oxygen engine is only about $3{,}300\ \text{K}$ — more than a thousand kelvin cooler than our estimate. Where did $1{,}400\ \text{K}$ of temperature go? (This is an order-of-magnitude illustration, Tier 3; real values need equilibrium chemistry software.)

The missing energy reveals a piece of physics that puts a hard cap on chemical rockets: dissociation. Above about $2{,}500\ \text{K}$, the product molecules do not sit quietly as water. They begin to tear apart — $\text{H}_2\text{O} \rightleftharpoons \text{H} + \text{OH}$, $\text{H}_2\text{O} \rightleftharpoons \text{H}_2 + \tfrac{1}{2}\text{O}_2$, and more — and every one of those bond-breakings absorbs energy, the same energy combustion just released. At rocket temperatures a substantial fraction of the products exists as these dissociated fragments (H, OH, O, free radicals), and the energy tied up in them is energy not available to raise the temperature. Dissociation acts as a chemical thermostat: push in more energy and the gas responds by dissociating more, not by getting proportionally hotter. This is why no combination of chemical reactants gets you much past $\sim 3{,}600\ \text{K}$ in practice, and it is a ceiling chemistry itself imposes.

🔗 Connection: Dissociation is not purely a loss. As the exhaust expands and cools in the nozzle (Chapter 19), those fragments can recombine — $\text{H} + \text{OH} \rightarrow \text{H}_2\text{O}$ — releasing their stored energy back into the flow and adding a little exhaust velocity. Whether they recombine in the fraction of a millisecond available is a genuine question that separates "frozen flow" from "equilibrium flow" analysis, and it is one of the reasons the real exit velocity takes a computer, not a formula, to predict exactly.

The second knob: molecular weight. Temperature is only half the story. Recall from kinetic theory that at a given temperature, lighter molecules move faster: the mean thermal speed of a gas molecule goes as $\bar{v} \propto \sqrt{T/\mathcal{M}}$, where $\mathcal{M}$ is the molar mass. A nozzle's job (Chapter 19) is to convert that random thermal motion into orderly directed motion, and it preserves the same scaling. The result — which we will derive from nozzle thermodynamics in the next chapter but state here because it is the organizing idea of this one — is the ideal exhaust velocity:

$$ v_e = \sqrt{\;\frac{2\gamma}{\gamma-1}\,\frac{R_u\,T_c}{\mathcal{M}}\left[\,1 - \left(\frac{p_e}{p_c}\right)^{(\gamma-1)/\gamma}\right]\;}, $$

where $R_u = 8.314\ \text{J/(mol·K)}$ is the universal gas constant, $\gamma$ is the ratio of specific heats of the exhaust (here $\gamma$ is the specific-heat ratio, not a flight-path angle — a typical hot-gas value is $\gamma \approx 1.2$), $\mathcal{M}$ is the mean molar mass of the exhaust mixture, and $p_e/p_c$ is the ratio of nozzle-exit to chamber pressure. The bracket and the $\gamma$ factors are set by the nozzle and are much the same for every propellant; strip them away and the propellant chemistry lives entirely in two numbers:

$$ \boxed{\; v_e \;\propto\; \sqrt{\dfrac{T_c}{\mathcal{M}}} \;} $$

🚪 Threshold Concept — the two knobs. Almost everything about propellant selection follows from this one proportionality. Exhaust velocity — and therefore specific impulse, and therefore, through the rocket equation, your entire vehicle — is governed by the ratio of flame temperature to exhaust molecular weight. There are only two ways for a chemist to give a rocket more $v_e$: make the fire hotter, or make the exhaust lighter. Once you internalize $v_e \propto \sqrt{T_c/\mathcal{M}}$, you can look at any propellant combination and predict, without a lab, roughly where its performance will land — and you will never again be surprised that the coldest-burning major fuel is also the most efficient.

That last remark is the payoff, and it deserves to be stated as the misconception it corrects.

⚠️ Common Misconception: "Hydrogen wins because it burns the hottest." It does not. A hydrogen–oxygen flame is actually cooler than a kerosene–oxygen or methane–oxygen flame at their normal operating points. Hydrogen wins on the other knob: its exhaust is by far the lightest. Burned to water and diluted with the leftover hydrogen that engines deliberately carry, hydrogen exhaust has a mean molecular weight around $\mathcal{M} \approx 13\ \text{g/mol}$, against roughly $\mathcal{M} \approx 23$ for kerosene. Because $v_e \propto \sqrt{T_c/\mathcal{M}}$, a propellant can lose on temperature and still win big on efficiency by having light exhaust. Hydrogen is the efficiency champion not because of how hot it burns but because of what it leaves behind.

Let us verify the claim numerically, because it is the intellectual center of the chapter.

Worked Example: hydrogen versus kerosene, by the two knobs. Take representative operating values (Tier 2, approximate): LOX/LH$_2$ at $T_c \approx 3{,}300\ \text{K}$, $\mathcal{M} \approx 13\ \text{g/mol}$; LOX/RP-1 at $T_c \approx 3{,}600\ \text{K}$, $\mathcal{M} \approx 23\ \text{g/mol}$. Kerosene burns about $300\ \text{K}$ hotter. Compare the knob: $$\sqrt{\frac{T_c}{\mathcal{M}}}\bigg|_{\text{LH2}} = \sqrt{\frac{3300}{13}} = \sqrt{254} = 15.9, \qquad > \sqrt{\frac{T_c}{\mathcal{M}}}\bigg|_{\text{RP-1}} = \sqrt{\frac{3600}{23}} = \sqrt{157} = 12.5.$$ The ratio is $15.9/12.5 = 1.27$: hydrogen should deliver about $27\%$ more exhaust velocity, despite the cooler flame. That tracks reality — vacuum $I_{sp}$ of about $450\ \text{s}$ for LOX/LH$_2$ versus about $340\ \text{s}$ for LOX/RP-1 is a ratio of $1.32$, the rest coming from chamber-pressure and nozzle differences. The light exhaust, not the hot flame, is doing the work.

We turn the boxed formula into a reusable tool in the Mission Design Checkpoint, and Chapter 19 will justify every factor in it. For now, hold onto the boxed proportionality: it is the lens for the rest of the chapter.

🔄 Check Your Understanding 1. Our naive energy balance predicted a hydrogen–oxygen flame near $4{,}700\ \text{K}$, but the real chamber sits near $3{,}300\ \text{K}$. Name the physical effect responsible and say which way it pushes. 2. Propellant A burns at $3{,}600\ \text{K}$ with exhaust $\mathcal{M} = 22$; propellant B burns at $3{,}200\ \text{K}$ with $\mathcal{M} = 14$. Without a calculator, which has the higher $v_e$, and why?

Answers

1. Dissociation: above ~2,500 K the products (water) partly tear into H, OH, O, and H$_2$, and those bond-breakings absorb energy, capping the temperature. It pushes the flame temperature down from the naive estimate. 2. Propellant B. Compare $T_c/\mathcal{M}$: A gives $3600/22 = 164$; B gives $3200/14 = 229$. B's lighter exhaust more than compensates for its cooler flame, so $\sqrt{T_c/\mathcal{M}}$ — and thus $v_e$ — is larger.

18.3 Propellant families

We can now understand why real engines burn the specific propellant combinations they do, and at the specific proportions they choose. The controlling dial is the mixture ratio.

Definition (mixture ratio). The mixture ratio $r$ (often written O/F) is the mass of oxidizer consumed per unit mass of fuel: $r = \dot m_{\text{ox}} / \dot m_{\text{fuel}}$. It is compared against the stoichiometric ratio: an engine runs fuel-rich when $r$ is below stoichiometric (excess fuel) and oxidizer-rich when it is above. Almost every engine runs deliberately fuel-rich.

Why fuel-rich? Two reasons, and the deeper one is pure $v_e \propto \sqrt{T_c/\mathcal{M}}$. Consider hydrogen–oxygen, and track the exhaust as we move away from the stoichiometric $r = 8$. Below stoichiometric, some hydrogen goes unburned and rides out in the exhaust. That unburned H$_2$ is the lightest molecule there is, so it drags the mean molecular weight down hard. In fact, for complete combustion to water plus leftover hydrogen, a clean piece of bookkeeping gives the mean molar mass as a simple function of mixture ratio:

$$ \mathcal{M}(r) = 2\,(1 + r) \quad \text{g/mol} \qquad (\text{H}_2/\text{O}_2,\ r \le 8,\ \text{frozen composition}). $$

(The derivation: per gram of hydrogen you have $0.5$ mol of H atoms' worth of H$_2$; the oxygen burns some to water and the rest of the hydrogen survives, and the total product always works out to $0.5$ mol carrying $1 + r$ grams — so $\mathcal{M} = (1+r)/0.5 = 2(1+r)$.) At stoichiometric $r = 8$ this gives the pure-water value $\mathcal{M} = 18$; drop to $r = 4$ and it falls to $\mathcal{M} = 10$. Now weigh that against the temperature, which falls as you go richer because less fuel is actually being oxidized:

The knob $\sqrt{T_c/\mathcal{M}}$ does not peak at the hottest, stoichiometric mixture — it peaks well fuel-rich, near $r \approx 3.5$, because the plunging molecular weight outruns the falling temperature until the temperature finally collapses. (The temperatures here are illustrative, Tier 3, folding in dissociation; the trend is what matters.) This is the analytical reason every hydrogen engine runs rich: the RS-25 at $r \approx 6$, upper stages richer still. They give up flame temperature on purpose to buy lighter, faster exhaust.

🔧 Engineering Reality: so why not run at the $\sqrt{T_c/\mathcal{M}}$ optimum near $r = 3.5$? Because a rocket is a system, not just a chamber. Running that rich means carrying enormous volumes of low-density liquid hydrogen, which demands bigger, heavier tanks — and that structural mass eats into the mass ratio the Isp was supposed to improve. The flown mixture ratio (RS-25 at ~6) is a compromise that sacrifices a few seconds of $I_{sp}$ to shrink the hydrogen tank. There is also a hard floor: run too rich and the flame temperature drops so far the reaction becomes sluggish and hard to sustain. The optimum on paper and the optimum in a vehicle are different numbers — a recurring lesson of Theme 4, mass is the enemy.

With mixture ratio in hand, here are the four families that fly, compared on the terms that matter. All performance and property numbers are approximate operating values (Tier 2); densities are bulk densities computed from component densities at the listed mixture ratio.

Read the table through the two knobs and the density column and the entire logic of engine design appears:

• LOX/LH$_2$ owns the top of the $I_{sp}$ chart entirely on the molecular-weight knob — its exhaust is half the weight of anyone else's. It pays in density: a bulk density near $0.36\ \text{g/cm}^3$ means its tanks are three times the volume of a kerosene stage's, so it is chosen where efficiency is carried a long way (upper stages, deep-space kick stages) rather than where thrust-per-volume rules (first stages).
• LOX/RP-1 is the dense, high-thrust workhorse. Its heavier exhaust caps its $I_{sp}$, but a bulk density near $1.0\ \text{g/cm}^3$ packs enormous energy into a small, structurally efficient first stage. Kerosene's drawback is chemical: burning a big carbon-chain molecule fuel-rich deposits soot and coke on injector and chamber walls, which fouls reusable hardware — a problem the next family sidesteps.
• LOX/CH$_4$ (methalox) is the modern compromise, and the reason this chapter's anchor vehicle exists. Its $I_{sp}$ beats kerosene by 20–40 seconds, its density beats hydrogen more than twofold, and — crucially — a single small carbon methane molecule burns cleanly, leaving little soot, so engines can be reused with minimal refurbishment.
• N$_2$O$_4$/UDMH trades performance for the ability to sit and wait. Its numbers are modest, but it is storable for years and hypergolic, which is why it flies on spacecraft that must reignite reliably after months in space.

🔗 Connection: the Starship anchor — why methane? Across this book we keep asking why SpaceX made the choices it did. Their most contrarian propulsion decision was methane, which no operational orbital rocket had ever used. The chapter's physics answers it: methane sits in the sweet spot of every trade at once — higher $I_{sp}$ than kerosene (lighter, partly hydrogen-bearing exhaust), far denser than hydrogen (compact tanks, one engine family for both stages), clean-burning (reusable without a soot-cleaning teardown), and, as 18.4 and 18.6 will show, storable near LOX temperature and manufacturable on Mars. No single property is best-in-class, but no rival wins on all four. We give Raptor and the full case study their due in Chapter 38.

📜 From History: Tsiolkovsky ordered the menu in 1903. The same deaf schoolteacher who wrote down the rocket equation (Chapter 3) also reasoned, purely from physics, that the best chemical propellants would be liquid hydrogen and liquid oxygen — decades before anyone could liquefy or handle them, and long before the molecular-weight argument was formalized. He could not have run our $\sqrt{T_c/\mathcal{M}}$ numbers, but he saw that the lightest fuel would give the fastest exhaust. Theory ran a half-century ahead of the hardware, again.

🐛 Find the Error. A student sizing an upper stage argues: "Kerosene–oxygen burns $350\ \text{K}$ hotter than hydrogen–oxygen, and hotter means faster exhaust, so I'll get more $I_{sp}$ from kerosene and a smaller, denser stage too. Kerosene upstairs is strictly better." Two things are wrong. What are they?

Answer

First, "hotter means faster exhaust" ignores the other knob. Exhaust velocity goes as $\sqrt{T_c/\mathcal{M}}$, and kerosene's exhaust ($\mathcal{M} \approx 23$) is nearly twice as heavy as hydrogen's ($\mathcal{M} \approx 13$); the molecular-weight penalty swamps the $350\ \text{K}$ temperature gain, so hydrogen delivers more $I_{sp}$, not less. Second, the reasoning treats $I_{sp}$ and density as if both favored kerosene, but they trade against each other: kerosene is denser, which is genuinely valuable — just on first stages, where propellant is burned quickly and low down. On an upper stage, whose propellant is carried nearly to orbit, the higher-$I_{sp}$ hydrogen usually wins the vehicle-level trade despite its bulk. The student has the density fact right and the efficiency conclusion exactly backwards.

18.4 Cryogenics and storability

Look again at the "type" column of the table. Whether a propellant must be kept frigid or can sit at room temperature is not a footnote — it reshapes the entire vehicle and the entire ground operation around it.

Definition (cryogenic propellant). A cryogenic propellant is one that is a gas at ordinary temperatures and must be liquefied and stored at very low temperature — conventionally below about $120\ \text{K}$ ($-153\,^\circ\text{C}$) — to be dense enough to use. Liquid oxygen boils at $90\ \text{K}$ ($-183\,^\circ\text{C}$), liquid methane at $112\ \text{K}$ ($-161\,^\circ\text{C}$), and liquid hydrogen at a ferocious $20\ \text{K}$ ($-253\,^\circ\text{C}$), only twenty degrees above absolute zero.

The opposite of cryogenic is storable: liquid at ordinary conditions and able to sit in a tank for months or years without refrigeration. RP-1 kerosene is storable; so are the hypergolics N$_2$O$_4$ and the hydrazine-family fuels. This spectrum — from hydrogen's $20\ \text{K}$ to room-temperature kerosene — drives a set of trade-offs that often matter more than a few seconds of $I_{sp}$.

Boiloff. A cryogen in a tank is perpetually absorbing heat from its surroundings, and it sheds that heat the only way it can: by evaporating. This continuous evaporative loss, called boiloff, means a cryogenic stage is quietly emptying itself from the moment it is filled. Hydrogen, with the largest temperature gap to the outside world and a tiny latent heat, boils off fastest; a hydrogen upper stage can lose a percent of its propellant per hour on the pad and cannot be left fueled for long. This is why cryogenic vehicles are loaded in the final hours of a countdown and why long-duration missions struggle to keep hydrogen. We treat boiloff as a propellant-management problem in Chapter 22, and the insulation that fights it — multilayer blankets, foam, and the physics of radiative and conductive heat leak — in Chapter 24.

Hydrogen's special miseries. Beyond boiloff, hydrogen is simply hard to contain. It is the smallest molecule, so it leaks through seals and even diffuses into metals, where it causes hydrogen embrittlement — weakening the very tanks and lines meant to hold it. Its extreme cold freezes air solid on any exposed surface, and its wide flammability range makes leaks dangerous. Every one of these is a reason a designer might accept a lower-$I_{sp}$ propellant for an easier life.

💡 Intuition: why methane is the "space-storable" Goldilocks. Liquid methane boils at $112\ \text{K}$ and liquid oxygen at $90\ \text{K}$ — a gap of only about twenty degrees. That closeness is a quiet superpower. The two propellants can share thermal-management hardware, sit behind a common insulated bulkhead, and even be gently sub-cooled together, whereas a hydrogen–oxygen vehicle must manage a $70\ \text{K}$ gulf between its two tanks. Methane's boiloff is mild, its density is high, and its storage temperature is close enough to oxygen's that a spacecraft can plausibly keep both through a months-long cruise. Hydrogen is the efficiency champion for a quick sprint to orbit; methane is the propellant you can actually live with on the way to Mars.

🔧 Engineering Reality: sub-cooling and densification. A clever trick blurs the line between "just barely liquid" and "densely liquid." By chilling a cryogen below its boiling point — sub-cooling it — you make it denser, so more mass fits in the same tank and the boiloff clock slows. Falcon 9 loads deeply chilled, densified LOX (and chilled RP-1) for exactly this reason, squeezing several percent more propellant into fixed tanks; the practice is one reason its late-load countdown is so tightly choreographed. Densification buys mass ratio — a direct assault on the rocket equation — at the cost of a much fussier ground operation.

The storability axis also explains an entire class of propellant that looks unimpressive on the $I_{sp}$ chart but is indispensable: the hypergolics.

Definition (hypergolic propellant). A hypergolic propellant is a fuel–oxidizer combination that ignites spontaneously on contact, with no spark, igniter, or external ignition source. The classic pairing is dinitrogen tetroxide (N$_2$O$_4$) with a hydrazine-family fuel — UDMH, monomethylhydrazine (MMH), or the 50/50 blend Aerozine-50. Bring the two liquids together and they simply catch fire, every time.

Hypergolics buy two things a hot cryogenic engine cannot easily offer: reliability of ignition and storability. There is no igniter to fail — a plumbing problem is the only failure mode, and the moment the valves open, the engine lights. Combined with room-temperature storage, that makes hypergolics the natural choice wherever an engine must wait a long time and then fire on command without fail: the Apollo Service Module engine that had to reignite to leave lunar orbit, the reaction-control thrusters that point nearly every spacecraft, and the upper stages of vehicles like Proton. The price is toxicity — these are among the nastiest industrial chemicals in use, requiring sealed suits and exclusion zones — which is precisely the problem the green propellants of 18.6 set out to solve.

🔄 Check Your Understanding 1. Rank LH$_2$, LCH$_4$, and RP-1 from hardest to easiest to store, and give the one-word reason for the extremes. 2. Why can a hypergolic-fueled spacecraft engine be trusted to reignite after six months in space when a cryogenic one generally cannot?

Answers

1. Hardest: liquid hydrogen (cryogenic at 20 K, fast boiloff, leaks, embrittlement). Middle: liquid methane (mildly cryogenic at 112 K, near LOX's temperature). Easiest: RP-1 kerosene (storable — liquid at room temperature). 2. Two reasons: hypergolics are storable at ordinary temperature, so nothing has boiled away or needs refrigeration over six months; and they ignite on contact, so there is no igniter or hot-start system that must survive the wait and work on the first try. Open the valves and the propellants light themselves.

18.5 Combustion instability

We come to the hardest problem in the chapter, and arguably in all of chemical propulsion. A combustion chamber is not merely hot; it is a resonant cavity full of the most violent chemistry humans sustain on purpose, and under the wrong conditions that chemistry and that cavity can lock together into an oscillation that destroys the engine faster than any sensor can react.

Definition (combustion instability). Combustion instability is a self-amplifying oscillation of chamber pressure and heat release, in which fluctuations in the combustion process couple to the acoustic (pressure) modes of the chamber so that each reinforces the other. Small perturbations, instead of damping out, grow — sometimes to pressure swings larger than the mean chamber pressure — with catastrophic heat-transfer spikes that can burn through chamber walls in milliseconds.

The physical heart: the Rayleigh criterion. Why should a fire and an echo feed each other? Lord Rayleigh saw the principle in 1878, studying singing flames in tubes. An oscillation gains energy when heat is added to the gas in phase with the pressure oscillation — that is, when the fire burns hardest at the instant the local pressure is highest. Formally, the oscillation grows when, integrated over a cycle,

$$ \oint p'\,\dot q'\,\,dt \;>\; 0, $$

where $p'$ is the fluctuating part of the pressure and $\dot q'$ the fluctuating part of the heat-release rate. The intuition is a playground swing: push in time with the swing's motion and the amplitude climbs; push out of phase and it dies. In a rocket chamber, a chance pressure ripple can momentarily improve mixing or injection, which makes the propellant burn a little harder, which raises the pressure further, which... If that feedback lands in phase with one of the chamber's natural acoustic modes, the ripple becomes a standing wave and the standing wave becomes a hammer.

Engineers sort the phenomenon by frequency, because different frequencies couple to different parts of the engine:

• Chugging (low frequency, ~10–200 Hz): the whole feed system breathes. A pressure rise in the chamber pushes back on the propellant lines, momentarily reducing flow, which drops the pressure, which lets flow surge back — a slow bulk oscillation coupling the chamber to its pumps and plumbing.
• Buzzing (intermediate, ~200–1,000 Hz): coupling to the injector and manifold dynamics; unpleasant and fatiguing but usually survivable.
• Screech / screaming (high frequency, >1,000 Hz, into several kHz): coupling to the acoustic resonant modes of the chamber itself — the tangential and radial pressure waves sloshing across the chamber. This is the killer. It concentrates heat transfer so violently that it can melt or rupture the chamber before a shutdown command completes.

📜 From History: the F-1 and the bombs. No engine tells this story better than the F-1, the giant kerosene engine of the Saturn V's first stage. Through the early 1960s the F-1 kept destroying itself in testing: high-frequency instability would erupt unpredictably and tear the chamber apart in well under a second, and no one could calculate in advance whether a given injector design would be stable. The Apollo schedule hung on solving it. The team's breakthrough was to stop trying to predict stability and instead test it directly: they detonated small explosive charges — literally bombs — inside a running F-1 chamber to slam it with a pressure pulse, then measured whether the engine damped the disturbance within a required time (a small fraction of a second). If it rang, the design failed; if it swallowed the blast and returned to smooth running, it passed. Hundreds of injector variants and thousands of tests later, a redesigned injector with baffles — radial and circumferential dividers standing off the injector face — broke up the transverse acoustic modes and made the engine reliably stable. It was one of the hardest and least glamorous engineering campaigns of the entire Moon program, and it was won empirically, exactly as the chapter's epigraph promises: one test result outweighed a thousand opinions.

The F-1's cure names the two families of fix still used today. Baffles physically obstruct the transverse sloshing of gas across the injector face, denying the dangerous modes a clean path. Acoustic cavities — small tuned resonators (Helmholtz cavities) recessed into the chamber wall near the injector — act as sinks that absorb energy at specific troublesome frequencies, the acoustic equivalent of a shock absorber. Injector design itself — the pattern, spacing, and impingement of the propellant jets — is the third and subtlest lever, shaping where and how fast heat is released so that the Rayleigh integral stays negative.

🔧 Engineering Reality: why this is still hard. Combustion instability couples chemistry (heat release on sub-millisecond timescales), acoustics (the chamber's geometry and speed of sound), and fluid dynamics (injection, atomization, mixing) into one nonlinear, ferociously sensitive problem. Small changes — a slightly different injector orifice, a new propellant temperature, a scaled-up chamber — can turn a rock-stable engine unstable. Modern computational fluid dynamics helps enormously, but no one yet fully predicts stability from first principles for a new engine. So the empirical discipline of the F-1 endures: serious engines are still bomb-tested and rated for how quickly they damp an imposed pulse. This is Theme 2 in its purest form. In space nothing can be repaired, and here the failure is measured in milliseconds — so the margin has to be designed and tested in, never assumed.

🧩 Productive Struggle. Before reading on: suppose an engine screeches at a particular high frequency tied to a tangential mode of its cylindrical chamber. You cannot change the propellants. Name two independent things you could change about the hardware to attack that specific mode, and say for each whether it removes the driving or adds damping.

One good answer

(1) Add injector baffles that physically interrupt the tangential (across-the-chamber) gas motion — this attacks the driving, denying that mode the coherent transverse flow it feeds on. (2) Cut tuned acoustic cavities into the chamber wall sized to resonate at the screech frequency — this adds damping, bleeding energy out of that specific mode. (A third: re-pattern the injector so peak heat release no longer lands in phase with the mode's pressure antinode — attacking the driving via the Rayleigh criterion directly.)

18.6 Green propellants and the future

The hypergolics that make spacecraft reliable also make them hazardous and expensive to handle: hydrazine and its cousins are acutely toxic and carcinogenic, and fueling a satellite with them means sealed protective suits, exclusion zones, and slow, costly operations. As launch cadence rises and reusability (Chapter 22) makes ground turnaround the bottleneck, the cost and risk of toxic propellant handling has become a real constraint. That pressure has produced a new class of propellant.

Definition (green propellant). A green propellant is one developed to reduce the toxicity and handling hazard of conventional propellants — chiefly hydrazine — while offering comparable or better performance and simpler, cheaper operations. "Green" here means primarily low toxicity and low handling cost, not carbon-free combustion; the aim is propellant a technician can work near without a hazmat suit.

The leading examples are ionic-liquid monopropellants — single premixed liquids that a catalyst decomposes to release energy, replacing toxic hydrazine monopropellant thrusters:

• ASCENT (formerly AF-M315E), a hydroxylammonium-nitrate blend, flew on NASA's Green Propellant Infusion Mission in 2019. It offers higher density and slightly higher performance than hydrazine, with far lower toxicity — you can handle it in ordinary protective gear.
• LMP-103S, an ammonium-dinitramide blend developed in Sweden, has flown operationally on spacecraft since the PRISMA mission. It, too, beats hydrazine on density-times-$I_{sp}$ while slashing handling cost.
• High-test hydrogen peroxide is enjoying a revival as a benign oxidizer and monopropellant: concentrated H$_2$O$_2$ decomposes over a catalyst into nothing but hot steam and oxygen.

Two honest caveats. First, these gains are mostly for small in-space thrusters and monopropellants, not for launch: the heavy lifting to orbit is still done by LOX with hydrogen, kerosene, or methane, and will be for the foreseeable future. Second, their specific impulse is only comparable to hydrazine's; the win is in density, safety, and operational cost, not in the two knobs of 18.2. Green propellants are a triumph of operability, which is exactly the currency reusability has made valuable.

🔗 Connection: methane as the quietly green choice, and the road to Mars. Methalox earns a place in this section too. A small, clean-burning molecule, methane leaves little soot and no toxic residue, and if the methane is synthesized rather than drilled, its lifecycle footprint falls further. Its decisive future credential, though, is off-world: methane and oxygen can be manufactured on Mars from the atmosphere's carbon dioxide and subsurface water ice, via the Sabatier reaction $$\text{CO}_2 + 4\,\text{H}_2 \;\rightarrow\; \text{CH}_4 + 2\,\text{H}_2\text{O},$$ paired with electrolysis of the water for oxygen. A vehicle that burns what it can make at its destination does not have to carry its return propellant across the solar system — a direct assault on the rocket equation. This in-situ resource utilization is a central reason Starship burns methane, and we develop it as part of the Mars architecture in Chapter 34.

Where does propulsion go from here? Chemistry's ceiling is fixed — we proved it in 18.1 and 18.2, and no fuel will break much past a vacuum $I_{sp}$ of ~465 seconds. The frontier therefore splits. For efficiency far beyond chemistry, you must leave chemistry entirely, trading thrust for enormous exhaust velocity in the electric thrusters of Chapter 20 and the nuclear and advanced concepts of Chapter 21. For getting off planets, where high thrust is non-negotiable, chemical combustion will reign for decades — but optimized relentlessly for reuse, clean operation, and manufacturability, with methane the current best answer and green propellants cleaning up everything that happens in space. The sustainability of what we launch, and leave behind, is a larger story we take up in Chapter 35.

🔄 Check Your Understanding 1. In one sentence, what does "green" actually refer to in a green propellant, and what does it not? 2. Why is the ability to make methane and oxygen on Mars a rocket-equation advantage, not merely a convenience?

Answers

1. It refers to low toxicity and low handling hazard/cost (a technician can work near it without a hazmat suit); it does not mean carbon-free or higher specific impulse — green monopropellants perform about like the hydrazine they replace. 2. Because a vehicle that manufactures its return propellant at the destination does not have to launch that propellant from Earth and haul it the whole way — and by the rocket equation, propellant you do not have to carry is mass ratio you do not have to pay for, on every leg before the last.

Mission Design Checkpoint: a propellant-choice note and propulsion.py

Your Mission Design Review now needs a propellant selection, and this chapter gives you the tools to justify one instead of guessing. Add a short propellant-choice note to your MDR that states, for your mission's main propulsion, which combination you have chosen and why — argued in the chapter's own terms.

The design decision. Pick from the four families of 18.3 and defend it against the trade-offs of 18.2–18.4:

• Track A (GEO comsat): the apogee/station-keeping propulsion runs for years and must reignite reliably after long coasts — storability and reliability dominate, favoring a hypergolic or a green monopropellant system; raw $I_{sp}$ is secondary.
• Track B (lunar lander): a short, high-thrust descent with restarts; a storable or methalox choice eases the long cruise and the need to reignite over the surface.
• Track C (Mars orbiter): a months-long cruise then orbit insertion — storables or space-storable methalox beat hydrogen's boiloff; if you are thinking ahead to a lander, methane's ISRU potential is a point in its favor.
• Track D (asteroid rendezvous): very long coasts and modest thrust make storables (or, looking to Chapter 20, electric propulsion) the natural fit.

Write one paragraph: your choice, the two or three properties that drove it (from $I_{sp}$, density, storability, restart, handling), and what you gave up.

The code. Extend the propulsion.py module (begun in Chapter 16) with the exhaust-velocity core of this chapter — the boxed formula of 18.2. Chapter 19 will wrap it with expansion-ratio optimization; here we use it to rank propellants.

# astrotools/propulsion.py  (Chapter 18 increment)
import math

G0 = 9.80665       # standard gravity, m/s^2
R_U = 8.314        # universal gas constant, J/(mol*K)

def exit_velocity(gamma, Tc, M, pe, pc):
    """Ideal exhaust velocity (m/s). Tc in K; M is exhaust molar mass in kg/mol;
    pe, pc are exit and chamber pressure in any shared unit. Nozzle theory: Ch. 19."""
    term = (2 * gamma / (gamma - 1)) * (R_U * Tc / M)
    bracket = 1 - (pe / pc) ** ((gamma - 1) / gamma)
    return math.sqrt(term * bracket)

# The chapter's central comparison: LOX/LH2 vs LOX/RP-1 (approximate operating values).
lh2 = exit_velocity(gamma=1.20, Tc=3300, M=0.013, pe=0.1, pc=200)   # light exhaust
rp1 = exit_velocity(gamma=1.22, Tc=3600, M=0.022, pe=0.1, pc=100)   # hotter, heavier

print(f"LOX/LH2 : ve = {lh2:.0f} m/s -> Isp = {lh2 / G0:.0f} s")
print(f"LOX/RP-1: ve = {rp1:.0f} m/s -> Isp = {rp1 / G0:.0f} s")
# Expected output:
# LOX/LH2 : ve = 4265 m/s -> Isp = 435 s
# LOX/RP-1: ve = 3278 m/s -> Isp = 334 s

The kerosene case burns $300\ \text{K}$ hotter yet lands a hundred seconds of $I_{sp}$ behind, entirely because its exhaust is heavier — the whole chapter, in two function calls. Feed your own chosen propellant's $(\gamma, T_c, \mathcal{M})$ into exit_velocity, get an $I_{sp}$, and carry it into the rocket equation from Chapter 3 to size your propellant load. This is the number that turns a propellant choice into a propellant mass in your MDR.

Summary

Combustion chemistry sets the ceiling on every chemical rocket. Carry these forward:

Numbers worth remembering: best chemical vacuum $I_{sp} \approx 450\text{–}465\ \text{s}$ (LOX/LH$_2$); stoichiometric H$_2$/O$_2$ is O/F $= 8$ (flown near 6); LOX boils at $90\ \text{K}$, LCH$_4$ at $112\ \text{K}$, LH$_2$ at $20\ \text{K}$; $v_e \propto \sqrt{T_c/\mathcal{M}}$.

Spaced Review

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

1. (§18.2, Ch. 17) Chapter 17 introduced engine cycles. A staged-combustion engine drives its turbines with a preburner running far off stoichiometric (very fuel- rich or very oxidizer-rich). Using this chapter's flame-temperature idea, why must a preburner avoid a near-stoichiometric mixture?
2. (Ch. 17) Name the three components a bipropellant liquid engine uses, in order, to turn two liquids into thrust: the part that mixes and sprays them, the part where they burn, and the part that accelerates the gas.
3. (Ch. 17) Why does full-flow staged combustion — which Raptor uses — let an engine reach higher chamber pressure than a gas-generator cycle that dumps its turbine-drive gas overboard?
4. (§18.3) In one sentence, why do essentially all engines run fuel-rich rather than at the hottest, stoichiometric mixture?

Answers

1. A near-stoichiometric preburner would reach the full ~$3{,}500\ \text{K}$ flame temperature, which no turbine blade can survive; running far off-ratio deliberately dilutes the combustion so the drive gas stays cool enough (well under ~$1{,}000\ \text{K}$) to spin the turbine without melting it. 2. The injector (mixes and atomizes the propellants), the combustion chamber (where they burn), and the nozzle (which accelerates the hot gas to exhaust velocity). 3. Because it burns all the propellant and routes the turbine-drive gas into the main chamber instead of dumping it, wasting no propellant and letting the pumps run harder — so it can sustain a higher chamber pressure (and thus higher performance) than an open cycle that throws its preburner exhaust away. 4. Because exhaust velocity goes as $\sqrt{T_c/\mathcal{M}}$, and running fuel-rich lowers the exhaust molecular weight faster than it lowers the temperature, raising $v_e$.

What's Next

We now know what sets the two numbers that matter — a flame temperature $T_c$ pinned by combustion chemistry and capped by dissociation, and an exhaust molecular weight $\mathcal{M}$ that the mixture ratio lets us tune. We even wrote down the formula that turns them into exhaust velocity. But we simply asserted that formula, borrowing its nozzle factors on faith. It is time to earn it. In Chapter 19 we derive the exhaust velocity from the thermodynamics of gas flowing through a converging–diverging nozzle: why a throat chokes at the speed of sound, why the diverging bell accelerates gas rather than slowing it, how the expansion ratio is tuned to altitude — and, at last, the full and rigorous reason that $v_e \propto \sqrt{T_c/\mathcal{M}}$, which is to say the complete reason hydrogen's light exhaust makes it the specific-impulse champion of every chemical rocket ever built.