Rocket staging

Why does a rocket throw away most of itself on the way up?

A rocket is one of the most expensive objects ever built, and its normal behaviour is to shed most of itself into the ocean within minutes of leaving the ground. No aircraft does this. No ship does this. The design looks like an admission of failure — so what is it an admission of? Something must make carrying the whole vehicle to orbit worse than destroying most of it. What could be so costly about carrying an empty tank?

Start with what a rocket actually is. It has no road, no air to push against, nothing outside itself to shove. It can only throw mass backwards and take the recoil. So its performance depends on two quantities and nothing else: how fast it throws things away, and how much of itself it has to throw.

Tsiolkovsky's rocket equation states the relationship exactly. The change in velocity a stage can achieve is

delta-v = v_e * ln(m_0 / m_f)

where v_e is the effective exhaust velocity, m_0 is the mass at ignition, and m_f is the mass at burnout. Read the shape of that before reading any numbers, because the shape is the whole story: the mass ratio sits inside a logarithm. Doubling the propellant does not double the speed gained. It adds one fixed increment — about 0.7 times the exhaust velocity — and to add that increment again you must double the whole vehicle again. Speed is bought in a currency that inflates viciously.

Now put numbers on it. Reaching low Earth orbit requires an orbital speed near 7.8 kilometres per second, plus roughly one and a half to two more spent fighting gravity and air on the way up — call the budget 9.4 kilometres per second. A good kerosene-and-oxygen engine has an effective exhaust velocity around 3 kilometres per second. So the required mass ratio is e raised to 9.4 divided by 3 — about twenty-three. Which means that at burnout, everything that is not propellant must be about four percent of what left the pad.

Four percent for tanks, engines, plumbing, avionics, structure — and the payload. Now ask what an empty propellant tank weighs. Even with excellent engineering, a stage's dry structure typically runs somewhere around five to ten percent of its own fuelled mass. The structure alone consumes the entire budget before a single kilogram of cargo is added. The vehicle cannot reach orbit. Not because we lack a clever trick, but because the logarithm has already spent the margin.

So what can be changed? Only two things appear in the equation. Raise the exhaust velocity: hydrogen and oxygen burn to give around 4.4 kilometres per second, which cuts the required ratio to roughly eight and a half — about twelve percent non-propellant. That is far more comfortable, except that liquid hydrogen is extremely low in density and must be kept near absolute zero, so its tanks are large and heavily insulated, and much of the gain is handed straight back as dry mass.

Or — and here is the move — change m_f. Notice that the equation punishes only the mass you are still carrying at burnout. An empty first-stage tank is dead weight that the second stage's engines would have to accelerate, for no return whatsoever. Drop it, and the remaining vehicle's mass ratio is computed fresh, from a much smaller starting mass, with its own smaller dry structure. Two stages each achieving 4.7 kilometres per second need a mass ratio under five apiece — entirely ordinary engineering. The same total delta-v, split, becomes achievable.

That is staging. It is not a way of getting more energy. It is a way of refusing to pay the logarithm's price on hardware that has finished its job.

A few refinements hold the reasoning together.

The equation says nothing about gravity, drag, or steering; those are the reason the real budget is 9.4 rather than 7.8 kilometres per second, and they are why a rocket wants to accelerate hard and turn horizontal early rather than hover. They are added to the requirement, not part of the equation.

Staging is not free, and the trade has a floor. Each additional stage brings its own engines, its own tanks, an interstage structure, and a separation event that must work exactly once. Beyond two or three stages, the added hardware mass and the added risk usually outweigh the shrinking benefit, which is why almost every orbital launcher settles at two or three rather than ten.

And a common misconception is worth naming: the discarded stage is not "wasted fuel". Its propellant was burned to useful effect and its structure did real work. What is discarded is a container whose remaining usefulness has become negative. Reusable boosters do not repeal any of this — they still separate and still stop accelerating the upper stage; they simply add landing propellant and legs, paying a performance penalty in exchange for recovering the hardware.

One honest caveat: single-stage-to-orbit is not proven impossible. With high-exhaust-velocity engines and extremely light structures the margin is thin but not obviously negative, and serious vehicles have been proposed. It has simply never been demonstrated, because staging wins comfortably at every level of technology so far achieved.

• Why launch sites cluster near the equator and fire eastward, and how much of the budget that saves.
• How the same equation governs interplanetary transfers, where gravity assists substitute for propellant.

Every term the collection defines is gathered in the glossary.