The Tsiolkovsky Rocket Equation

The equation and its symbols

In its ideal form — a single rocket firing in gravity-free, drag-free space — the Tsiolkovsky rocket equation is:

Δv = v_e · ln(m₀ / m_f)

Every symbol carries a precise meaning and unit:

• Δv — the ideal velocity change delivered by the burn, in m/s. Also called the vehicle's "delta-v budget."
• v_e — the effective exhaust velocity, in m/s: the momentum-averaged speed of the exhaust relative to the rocket, including the small pressure-thrust contribution at the nozzle exit. Related to specific impulse by v_e = I_sp · g₀.
• m₀ — the initial (wet) mass at ignition, in kg: structure + payload + propellant.
• m_f — the final (dry) mass at burnout, in kg: structure + payload only, all usable propellant spent.
• m₀/m_f — the dimensionless mass ratio R. The logarithm is natural (base e).

Because ln(1) = 0, a rocket that burns no propellant produces no Δv. And because the logarithm grows so slowly, wringing large Δv out of the equation requires the mass ratio to grow enormously — the exponential penalty at the heart of rocketry.

Where it comes from: momentum of a variable-mass body

A rocket is not a fixed lump of matter — it continuously ejects mass. The equation follows from conservation of momentum applied to the rocket-plus-exhaust system, which is closed even though the rocket alone is not.

At time t the rocket has mass m and velocity v, so momentum m·v. In an interval dt it ejects a small mass |dm| (note dm < 0, the rocket's mass decreases) at exhaust velocity v_e relative to itself — i.e. at velocity (v − v_e) in the inertial frame. Conserving total momentum in force-free space:

m·v = (m + dm)·(v + dv) + (−dm)·(v − v_e)

Expanding, discarding the second-order product dm·dv, and simplifying gives the differential rocket equation:

m·dv = −v_e·dm     ⇒     dv = −v_e · (dm/m)

The quantity −v_e·(dm/dt) is the thrust: force equals exhaust velocity times mass-flow rate, F = v_e·ṁ. Integrating dv from 0 to Δv as m falls from m₀ to m_f, with v_e treated as constant:

Δv = −v_e · ∫(dm/m) from m₀ to m_f = v_e · ln(m₀/m_f)

The logarithm is not a modelling choice — it is the exact integral of 1/m. It appears because each parcel of propellant must accelerate not only the payload but also all the propellant still sitting in the tanks above it.

Specific impulse and exhaust velocity

Engineers rarely quote v_e directly; they quote specific impulse, I_sp, the impulse delivered per unit weight of propellant consumed. In the weight-based convention it has units of seconds:

I_sp = v_e / g₀      ⇔      v_e = I_sp · g₀      (g₀ = 9.80665 m/s²)

The g₀ here is standard gravity as a unit conversion, not the local gravitational field — it is a historical convention that lets engineers using pounds-force and those using newtons agree on the same number of seconds. Higher I_sp means more Δv per kilogram of propellant. The trade is fundamental: chemical rockets get huge thrust but modest v_e; electric thrusters get tiny thrust but enormous v_e.

Propulsion system	Typical I_sp (s)	v_e ≈ I_sp·g₀ (m/s)
Solid rocket booster (Shuttle SRB)	~250	~2,450
Kerosene/LOX (Merlin 1D, vacuum)	~311	~3,050
Hydrazine monopropellant (attitude)	~220	~2,160
Hydrogen/LOX (RL10, RS-25)	~450	~4,400
Hall-effect / gridded-ion thruster	1,500–4,000+	15,000–40,000+
Nuclear thermal (NERVA-class)	~900	~8,800
Photon rocket (theoretical limit)	≈ 3.06 × 10⁷	c ≈ 3 × 10⁸

The exponential penalty of the mass ratio

Rearranging the equation for mass ratio shows the tyranny directly:

m₀/m_f = exp(Δv / v_e)

The required mass ratio grows exponentially with the ratio Δv/v_e. Propellant fraction — the share of liftoff mass that is fuel — is ζ = 1 − m_f/m₀ = 1 − exp(−Δv/v_e). With a good hydrogen/LOX exhaust velocity of v_e ≈ 4,400 m/s, watch how fast the tanks have to grow:

Target Δv	Δv / v_e	Mass ratio m₀/m_f	Propellant fraction
2.2 km/s	0.50	1.65	39%
4.4 km/s	1.00	2.72	63%
6.6 km/s	1.50	4.48	78%
9.4 km/s (LEO)	2.14	8.5	88%
13.2 km/s	3.00	20.1	95%
17.6 km/s	4.00	54.6	98%

To reach LEO on a single hydrogen stage you would need about 88% of your liftoff mass to be propellant — leaving only 12% for tanks, engines, plumbing, avionics, and payload. Real structural mass fractions are typically 6–10%, which leaves almost nothing for payload. This is why a single-stage-to-orbit (SSTO) chemical rocket is right at the edge of physical possibility, and why nearly every launcher ever flown has been staged.

Why staging beats a single stage

Dead structure — empty tanks and spent engines — is included in m_f, and m_f sits in the denominator of the mass ratio. Carrying an empty first-stage tank all the way to orbit means its mass counts against every kilogram of Δv the upper stages must still produce. Staging solves this by discarding empty hardware midflight so each subsequent stage begins with a fresh, favorable mass ratio.

Because Δv is additive across stages, the total is simply the sum:

Δv_total = Σ v_e,i · ln(m₀,i / m_f,i)

Consider a concrete comparison. Suppose each stage has v_e = 4,400 m/s and a structural fraction (dry-structure mass ÷ stage propellant+structure mass) of 8%. A single stage can, at best, approach Δv = v_e·ln(1/0.08) ≈ 4,400 × 2.53 ≈ 11.1 km/s only if it carried zero payload — impossible in practice, because a real payload adds to m_f. Split the same job into two stages and each stage's structural burden is paid separately: the second stage no longer has to haul the first stage's empty tank, so the achievable payload-carrying Δv climbs well past 9.4 km/s with margin. Three stages do better still. The Saturn V used three; the Falcon 9 uses two plus payload. Staging is the engineering answer to a logarithm.

Worked example: how much fuel to reach orbit

Take a vehicle with hydrogen/LOX upper propulsion, v_e = 4,400 m/s, that must deliver Δv = 9,400 m/s (LEO with losses). The mass ratio is:

m₀/m_f = exp(9400 / 4400) = exp(2.136) ≈ 8.47

If the payload we want in orbit is 20,000 kg and we optimistically assume all non-payload dry mass is negligible (so m_f ≈ 20,000 kg), then m₀ ≈ 8.47 × 20,000 ≈ 169,000 kg, of which about 149,000 kg is propellant. Now swap to a lower-performance kerosene/LOX exhaust velocity of v_e = 3,050 m/s:

m₀/m_f = exp(9400 / 3050) = exp(3.082) ≈ 21.8

The same 20,000 kg now needs m₀ ≈ 436,000 kg — more than 2.5× the liftoff mass, and over 400,000 kg of propellant. Dropping v_e by 30% didn't cost 30% more fuel; it cost more than double, because v_e sits in an exponent. This single calculation explains the relentless pursuit of high-I_sp engines.

Ideal Δv versus real-world velocity gain

The pure equation ignores gravity and the atmosphere. A launch vehicle actually experiences:

v_gained = Δv_ideal − Δv_gravity − Δv_drag − Δv_steering

• Gravity loss ≈ ∫ g·sin θ dt — thrust wasted holding the vehicle up while it climbs; typically 1.5–2 km/s for an Earth ascent. This is why rockets pitch over early (a "gravity turn") to spend as little time as possible fighting straight down.
• Drag loss ≈ ∫ (D/m) dt — aerodynamic resistance in the lower atmosphere; usually 0.3–0.5 km/s.
• Steering loss — thrust vectored off the velocity vector to control the trajectory.

So the ~7.8 km/s orbital velocity of LEO becomes a ~9.4 km/s ideal Δv budget once losses are added. Engineers size a vehicle by its Δv budget and require it to exceed the mission need plus all losses. The rocket equation still governs how much propellant that budget costs — losses only raise the target Δv.

Δv budgets for common missions

Maneuver	Approx. Δv (km/s)	Note
Earth surface → LEO	9.3–9.5	Includes gravity + drag losses
LEO → geostationary transfer (GTO)	2.4	First Hohmann burn
GTO → geostationary (GEO)	1.5	Circularize + plane change
LEO → trans-lunar injection	3.1	Toward the Moon
LEO → Mars transfer	3.6	Minimum-energy departure
Earth surface → escape (from rest)	~11.2	Ignoring losses; escape velocity
Moon surface → lunar orbit	~1.9	No atmosphere, ⅙ g

These budgets add along a mission. A probe to Mars pays LEO first, then trans-Mars injection, then capture — and each leg's propellant cost compounds through the exponential mass ratio. See escape velocity and the Hohmann transfer for the orbital-mechanics side of these numbers.

JavaScript — rocket-equation calculations

const g0 = 9.80665; // standard gravity, m/s^2 (unit conversion, not local g)

// Ideal delta-v from exhaust velocity and mass ratio
function deltaV(v_e, m0, mf) {
  return v_e * Math.log(m0 / mf);
}

// Effective exhaust velocity from specific impulse (seconds)
function exhaustVelocity(Isp) {
  return Isp * g0;
}

// Required mass ratio for a target delta-v
function massRatio(dv, v_e) {
  return Math.exp(dv / v_e);
}

// Propellant fraction (share of wet mass that is fuel)
function propellantFraction(dv, v_e) {
  return 1 - Math.exp(-dv / v_e);
}

// Hydrogen/LOX to LEO
const veLH2 = exhaustVelocity(450);                 // ~4413 m/s
console.log(`v_e (LH2/LOX): ${veLH2.toFixed(0)} m/s`);
console.log(`Mass ratio for 9400 m/s: ${massRatio(9400, veLH2).toFixed(2)}`); // ~8.4
console.log(`Propellant fraction: ${(propellantFraction(9400, veLH2)*100).toFixed(1)}%`); // ~88%

// Same job on kerosene/LOX costs far more fuel
const veRP1 = exhaustVelocity(311);                 // ~3050 m/s
console.log(`Mass ratio (RP-1/LOX): ${massRatio(9400, veRP1).toFixed(2)}`);   // ~21.8

// Staged total delta-v is the sum of each stage
function stagedDeltaV(stages) {
  // stages: [{ v_e, m0, mf }, ...]
  return stages.reduce((sum, s) => sum + deltaV(s.v_e, s.m0, s.mf), 0);
}

const twoStage = stagedDeltaV([
  { v_e: 3050, m0: 500000, mf: 130000 }, // stage 1
  { v_e: 4400, m0: 120000, mf: 30000 }   // stage 2
]);
console.log(`Two-stage total Δv: ${(twoStage/1000).toFixed(2)} km/s`); // ~10.21 km/s

// Ion thruster: tiny thrust, huge v_e -> huge delta-v from little propellant
const veIon = exhaustVelocity(3000);                // ~29420 m/s
console.log(`Ion Δv (m0/mf = 1.5): ${deltaV(veIon, 1.5, 1).toFixed(0)} m/s`); // ~11930 m/s

A short history

Konstantin Eduardovich Tsiolkovsky (1857–1935), a self-taught, largely deaf Russian schoolteacher, derived the equation and published it in his 1903 paper "Exploration of Cosmic Space by Means of Reaction Devices". Working in near-isolation in Kaluga, he grasped that reaction propulsion — not wings or cannons — was the only way to leave Earth, and that liquid propellants and multi-stage "rocket trains" would be required. He never built a rocket, yet his equation defined the field.

The result was independently rediscovered by others: Robert Goddard in the United States (who flew the first liquid-fuel rocket in 1926) and Hermann Oberth in Germany. The same 1/m integral had, in fact, been noted even earlier by William Moore (1813) and Russian artilleryman I. V. Meshchersky in his general variable-mass dynamics — but Tsiolkovsky's name is attached because he framed it explicitly for spaceflight. The equation underwrote every launch since: from the R-7 that lofted Sputnik to the Saturn V, the Space Shuttle, and today's reusable boosters.

Common misconceptions

• "The rocket pushes against the air/ground." No — thrust is pure momentum exchange with the exhaust. A rocket works better in vacuum, where there is no back-pressure at the nozzle and no drag. Newton's third law acting on the ejected mass is the whole story.
• "Δv equals final speed." Δv is a capability — the total velocity change the propellant can buy. Actual speed gained is Δv minus gravity, drag, and steering losses, and depends on the trajectory.
• "Double the Δv, double the fuel." The relationship is exponential, not linear. Doubling Δv squares the mass ratio, because m₀/m_f = exp(Δv/v_e).
• "g₀ in I_sp is the local gravity." It isn't — g₀ = 9.80665 m/s² is a fixed unit-conversion constant. A rocket's I_sp doesn't change if you fly it to the Moon.
• "F = ma applies to the rocket as-is." The rocket has variable mass, so you must use the momentum form; naïvely writing F = m·(dv/dt) with changing m and no thrust term gives the wrong dynamics. The correct thrust is F = v_e·ṁ.
• "Staging is just about dropping weight." It's specifically about resetting the mass ratio: each stage's structural mass stops penalizing the stages above it once jettisoned.