Visual walkthrough — Nozzle area ratio ε = A_e - A - — choosing for optimal performance

---

Step 1 — What a nozzle is, before any symbols

WHAT. A nozzle is a pipe that first squeezes, then flares. Hot gas enters slow and at very high pressure on the left; it leaves fast and thin on the right. The narrowest slice in the middle has a special name — we will call its area $A^{\ast }$ (read "A-star"). The mouth at the far right, where gas escapes into the sky, has area $A_e$ (the subscript $e$ just means "exit").

WHY these two areas. Everything on this page is a competition between what happens at the throat and what happens at the exit. Their ratio is the one number we can choose when we build the nozzle:

PICTURE. Follow the outline: it pinches to $A^{\ast }$, then opens out to $A_e$. The gas arrow speeds up left-to-right.

---

Step 2 — The two ways gas pushes the rocket

WHAT. Thrust — the forward shove on the rocket — comes from two separate effects, and we must name both.

1. Momentum thrust. Every second, a mass $\dot{m}$ (read "m-dot", kilograms of gas per second) leaves at speed $v_e$ (exit velocity). Throwing mass backward shoves the rocket forward: this contributes $\dot{m}{v}_e$.
2. Pressure thrust. The gas at the exit still presses outward with pressure $p_e$. The outside air presses inward with pressure $p_a$ (ambient). Over the mouth area $A_e$ the net push is $(p_e-p_a)A_e$.

WHY it matters here. Notice $p_e$ and $A_e$ both change when we open the flare. That coupling is the whole story. Look at the two coloured arrows: one is the mass-throwing jet, the other is the pressure difference smeared over the mouth.

---

Step 3 — The tug-of-war: flare more, and two things fight

WHAT. Suppose we make the nozzle flare a little more — we add a thin ring of extra area $d{A}_e$ (the "$d$" means "a tiny bit of"). Because the throat stays choked (Step 1), $\dot{m}$ does not change; only what happens downstream shifts. Two things happen at once:

• The gas gets a little more room, so it speeds up: $v_e$ rises → momentum thrust rises. Good.
• The gas spreads thinner, so its pressure $p_e$ drops. Once $p_e$ falls below $p_a$, the pressure term turns negative — the outside air now pushes back. Bad.

WHY differentiate. We can't eyeball the winner, so we ask calculus the exact question: "If I add area $d{A}_e$, does total thrust $F$ go up or down?" The tool for "does this rise or fall as I nudge that" is the derivative $dF/d{A}_e$ — the slope of thrust versus flare. When the slope is zero, we are at the peak.

PICTURE. The added ring $d{A}_e$ sits at the lip. The green arrow (velocity gain) and the amber arrow (pressure loss) pull the total in opposite directions.

---

Step 4 — Balancing the forces on one wall slice

WHAT. With $\dot{m}$ fixed, the momentum-thrust part of $F$ is $\dot{m}{v}_e$ and the pressure part is $(p_e-p_a)A_e$. Nudge the flare by $d{A}_e$ and take the total differential of $F=\dot{m}{v}_e+(p_e-p_a)A_e$:

$dF=\dot{m}d{v}_e+d(p_e{A}_e)-p_{a}d{A}_e$

• $\dot{m}d{v}_e$ = extra momentum thrust from the speed-up.
• $d(p_e{A}_e)$ = change in the exit pressure force (area grew, pressure shrank).
• $-p_{a}d{A}_e$ = extra push-back from air on the new ring (ambient $p_a$ is constant, so only $A_e$ varies).

WHY the internal terms collapse — the two-line bridge. First expand the product using the product rule:

$d(p_e{A}_e)=A_{e}d{p}_e+p_{e}d{A}_e$

So the two internal terms are

$\dot{m}d{v}_e+d(p_e{A}_e)=\dot{m}d{v}_e+A_{e}d{p}_e+p_{e}d{A}_e.$

Now the key physics: the steady 1-D momentum equation for the gas flowing through the slice says that the force accelerating the gas is the pressure force on it, $\dot{m}d{v}_e=-A_{e}d{p}_e.$ (Read it as: the gas speeds up exactly because pressure falls across the slice — momentum gained equals pressure force applied.) Substitute this in and the $A_{e}d{p}_e$ pieces cancel:

$\underset{\dot{m}d{v}_e}{\underbrace{-A_{e}d{p}_e}}+A_{e}d{p}_e+p_{e}d{A}_e=p_{e}d{A}_e.$

That is the whole collapse: the velocity gain and the pressure-times-area change are not independent — they are opposite sides of the same momentum balance, and what survives is just the wall pressure $p_e$ acting over the new ring.

PICTURE. On the flaring wall, the gas presses out with $p_e$ over the new ring, the air presses in with $p_a$ over the same ring. Net = the difference.

---

Step 5 — The optimum falls out (and it is a true maximum)

WHAT. Put the collapsed result into the $dF$ expression:

$dF=\underset{\text{gas pushes out}}{\underbrace{p_{e}d{A}_e}}-\underset{\text{air pushes in}}{\underbrace{p_{a}d{A}_e}}=(p_e-p_a)d{A}_e$

Divide by $d{A}_e$ to get the slope of thrust versus flare:

$\frac{dF}{d{A}_e}=p_e-p_a$

WHY set it to zero. A hilltop is where the slope is flat. Thrust is maximised when adding more area neither helps nor hurts:

$\frac{dF}{d{A}_e}=0⟹\boxed{p_e=p_a}$

WHY it is a maximum, not a minimum. A zero slope alone could be a valley bottom. We must check the second derivative $d^{2}F/d{A}_e^2$ — the rate at which the slope itself changes. Differentiate $dF/d{A}_e=p_e-p_a$ again (recall $p_a$ is fixed):

$\frac{d^{2}F}{d{A}_e^2}=\frac{d{p}_e}{d{A}_e}.$

In the diverging, supersonic section, opening more area always lowers the exit pressure, so $d{p}_e/d{A}_e<0$. Therefore

$\frac{d^{2}F}{d{A}_e^2}=\frac{d{p}_e}{d{A}_e}<0\text{at }{p}_e=p_a,$

which is exactly the condition for a maximum (curve bending downward). Equivalently, the slope $p_e-p_a$ goes from $+$ (while $p_e>p_a$) through $0$ to $-$ (once $p_e<p_a$): a sign change from positive to negative confirms the peak.

• If $p_e>p_a$: slope positive → still climbing → under-expanded, nozzle too short.
• If $p_e<p_a$: slope negative → past the peak → over-expanded, nozzle too long (and risking Flow Separation in Over-expanded Nozzles).
• If $p_e=p_a$: flat top with downward bend → maximum thrust.

PICTURE. A hill: $F$ climbs while $p_e>p_a$, peaks where the two pressures meet, then falls as $p_e<p_a$.

---

Step 6 — Turning a pressure into a shape (the geometry $\epsilon$)

WHAT. The rule $p_e=p_a$ tells us the pressure we want — but we build a shape. We need the bridge from "desired exit pressure" to "how wide to flare," i.e. to $\epsilon$.

We use $\gamma$ (ratio of specific heats, a gas property), $M_e$ (exit Mach number — exit speed divided by the local speed of sound), and $p_0$ (the chamber stagnation pressure, i.e. the high pressure of the gas at rest inside the combustion chamber).

(1) Pressure sets the exit Mach. $\frac{p_0}{p_e}={\left(1+\frac{\gamma -1}{2}{M}_e^2\right)}^{\frac{\gamma }{\gamma -1}}$

• $p_0$ = chamber stagnation pressure (defined just above).
• $p_e$ = exit pressure we chose to equal $p_a$.
• A bigger pressure drop $p_0/p_e$ forces a bigger $M_e$.

(2) Mach sets the area ratio. $\epsilon =\frac{A_e}{A^{\ast }}=\frac{1}{M_e}{\left[\frac{2}{\gamma +1}\left(1+\frac{\gamma -1}{2}{M}_e^2\right)\right]}^{\frac{\gamma +1}{2(\gamma -1)}}$

WHY the $1/M_e$ out front matters. Because of that factor, the curve of $\epsilon$ versus $M$ dips to a minimum of $1$ at $M=1$ (the throat) and rises on both sides. So one $\epsilon >1$ matches two Mach numbers — one subsonic, one supersonic. Downstream of the throat the flow is supersonic, so we always take the $M>1$ branch.

PICTURE. The U-shaped curve $\epsilon (M)$: minimum at the throat, two arms. A horizontal line at our chosen $\epsilon$ cuts it twice; we mark the supersonic (right) intersection.

---

Step 7 — Every case on one altitude ladder

WHAT. A built nozzle has a fixed $\epsilon$, so it expands gas to one fixed $p_e$ regardless of altitude. But $p_a$ falls as the rocket climbs. So the same nozzle passes through all three regimes:

• Low altitude ($p_a$ high, $p_a>p_e$): over-expanded, pressure term subtracts, separation risk.
• Design altitude ($p_a=p_e$): perfect, maximum thrust.
• High altitude / vacuum ($p_a$ low, $p_a<p_e$): under-expanded, gas keeps pushing but wastes some energy.

WHY it drives design. Sea-level stages take small $\epsilon$ (5–15); upper stages take large $\epsilon$ (tens–hundreds) because near-vacuum demands a tiny matched $p_e$. The device that dodges the whole trade-off is the aerospike.

PICTURE. A rocket rising through three atmosphere bands; the fixed $p_e$ jet compared against shrinking $p_a$ at each band.

---

The one-picture summary

PICTURE. Left: the nozzle with $A^{\ast }$ and $A_e$. Middle: the tug-of-war (velocity up, pressure down). Right: the thrust hill peaking at $p_e=p_a$. This single figure is the whole derivation.

Recall Feynman: the whole walkthrough in plain words

A nozzle is a squeeze-then-flare pipe. It shoves the rocket two ways: by hurling gas out the back (fast gas = more shove), and by the leftover gas-pressure pressing on the wide mouth against the outside air. Now play a game: open the flare a tiny bit more. The gas gets more room so it speeds up — good, more hurling-shove. But it also spreads thin, so its pressure drops. As long as the gas is still pushing harder than the outside air, opening more helps. The instant the gas pressure sinks to match the outside air, you're on top of the hill — open any more and the outside air starts winning and drags you back. So the best flare is the one where inside-pressure exactly equals outside-pressure, $p_e=p_a$. Since the outside air is thick near the ground and thin up high, the "best flare" is different at every altitude — which is why ground engines wear small flares and space engines wear enormous ones.

---

Connections

• Thrust Equation and Effective Exhaust Velocity
• De Laval Converging-Diverging Nozzle
• Isentropic Flow Relations
• Choked Flow and the Throat Condition
• Flow Separation in Over-expanded Nozzles
• Altitude Compensation — Aerospike Nozzles
• Specific Impulse Isp