Relativistic Rocket Equation

Here is a number that ends an argument. To push a chemical rocket — the good kind, liquid hydrogen and liquid oxygen, exhaust velocity $4.5\ \mathrm{km/s}$ , the engine that flew the Saturn V — to half the speed of light, you would need a mass ratio of

\frac{m_0}{m} \;\approx\; 10^{15{,}893}.

The observable universe contains something like $10^{80}$ atoms. The number above is not "difficult", not "a stretch", not "beyond current technology". It is a number with fifteen thousand eight hundred and ninety-four digits, and there is no arrangement of the matter in the universe that pays it. That is a fact about exhaust velocity, and this article explains exactly where it comes from — but the deeper fact is more interesting. Newton's rocket equation, $\Delta v = u\ln(m_0/m)$ , says velocity accrues logarithmically but without bound: burn enough fuel and you exceed $c$ . Relativity does not permit that, and the way it forbids it is beautiful and specific.

The rocket does not buy velocity. It buys rapidity. Every squirt of exhaust gives the ship a small velocity boost in its own instantaneous rest frame — and those boosts do not add as velocities, because velocities do not add. They add as rapidities, $\varphi = \operatorname{artanh}\beta$ , which is precisely the variable that does add under successive boosts. So the thing that grows linearly with $\ln(m_0/m)$ is not $v$ . It is $\varphi$ :

\varphi = \frac{u}{c}\,\ln\frac{m_0}{m}, \qquad\text{and then}\qquad \beta = \tanh\varphi.

And $\tanh$ of anything is less than one. You may burn the entire universe for fuel and $\tanh$ will still return a number below 1. The speed limit is not enforced by some new force that appears near $c$ . It is enforced by the shape of a hyperbolic tangent.

Key takeaways

Signature $(+,-,-,-)$ ; $P\cdot P = +m^2c^2$ . $m$ is the rocket's invariant rest mass and it decreases because the ship is throwing mass away — nothing here gains mass by moving.
The result, and the thesis in one line:

\boxed{\;\varphi = \frac{u}{c}\,\ln\frac{m_0}{m}\;} \qquad\text{and}\qquad \boxed{\;\beta = \tanh\varphi = \tanh\!\Big(\frac{u}{c}\ln\frac{m_0}{m}\Big)\;}

where $u$ is the exhaust speed relative to the rocket and $\varphi = \operatorname{artanh}\beta$ is rapidity.

Why rapidity: each burn is a small boost in the ship's own instantaneous rest frame. Velocities do not add across boosts; rapidities do (Rapidity and Hyperbolic Boosts owns that theorem). A rocket is a chain of boosts, so it accumulates rapidity, linearly — and $\varphi$ is exactly what the logarithm of the mass ratio buys.
The speed limit is free. $\beta = \tanh\varphi < 1$ for every finite $\varphi$ , and $\varphi$ is finite for every finite mass ratio. No fuel supply reaches $c$ , and nothing had to be added to the theory to make that so.
Newton is the $c\to\infty$ limit. For $\varphi \ll 1$ , $\tanh\varphi \approx \varphi$ , so $v \approx u\ln(m_0/m)$ — Tsiolkovsky, 1903, recovered exactly.
Invert it: $\dfrac{m_0}{m} = \Big(\dfrac{1+\beta}{1-\beta}\Big)^{c/2u}$ . This is the Doppler factor $k$ raised to the power $c/u$ — one formula, and the photon rocket is just the case $u = c$ .
Photon rocket ( $u = c$ , the theoretical optimum): $\dfrac{m_0}{m} = \sqrt{\dfrac{1+\beta}{1-\beta}} = k$ . To reach $0.99c$ : mass ratio 14.1. Achievable in principle.
Everything slower is brutal. A fusion drive at $u = 0.1c$ needs $3.1\times10^{11}$ for the same $0.99c$ . A chemical rocket needs $10^{15{,}893}$ for a mere $0.5c$ .
Key idea: the rocket buys rapidity, not velocity. Rapidity is unbounded; velocity is its hyperbolic tangent, and therefore is not.

The core insight

Why can a rocket not simply keep accelerating past $c$ ? Not because it gets heavier — it does not; its rest mass is an invariant, and anyway it is losing mass. The real answer is a statement about composition. In the ship's own instantaneous rest frame, each burn does exactly what a Newtonian burn does: it changes the ship's velocity by $\mathrm{d}v' = -u\,\mathrm{d}m/m$ , and there is no relativistic correction to that at all. The relativity is entirely in how those increments combine. A hundred boosts of $0.01c$ each do not make $c$ ; they make $\tanh(100 \times 0.01) = \tanh(1) = 0.762c$ . The quantity that adds across boosts is the rapidity, and the rocket's ledger — one increment per unit of $\ln(m_0/m)$ — is a ledger in rapidity. Velocity is what you get by taking a hyperbolic tangent at the end, and a hyperbolic tangent is bounded by one.

Where this lives today

Nowhere, in one sense: no rocket ever built has needed a relativistic correction, and none will for a very long time. The Saturn V's rapidity was $4\times10^{-5}$ , and the correction to its velocity was around a nanometre per second. But the equation is not idle, because it settles what is possible, and it settles it before any engineering begins. The exponent $c/u$ in $m_0/m = k^{c/u}$ is a verdict, not a difficulty: it says that interstellar flight at relativistic speed requires an exhaust velocity within an order of magnitude of $c$ , and that everything else is excluded by arithmetic rather than by engineering. That is why every serious proposal for interstellar propulsion either has a near- $c$ exhaust — antimatter annihilation, fusion products, a photon drive — or carries no fuel at all. The second option is the interesting one: a light sail, or a beam-pushed sail, sidesteps the rocket equation entirely by leaving the reaction mass at home. The Breakthrough Starshot concept is exactly this — a gram-scale sail driven by a ground-based laser — and the reason it exists is that the rocket equation's exponent leaves no other door open. And the physics content — that a chain of boosts accumulates rapidity — is not about rockets at all. It is the same statement as the fact that no sequence of accelerations reaches $c$ , and it is the cleanest demonstration in the subject that the speed limit is a property of how velocities compose, not a property of any force, engine, or mass.

The physical picture

Sit in the rocket. At this instant you are at rest — in your frame, which is as good as anyone's. You fire the engine for a moment and shove a bit of exhaust out the back at speed $u$ (measured by you, in your ship). By momentum conservation you recoil forward a little, by an amount $\mathrm{d}v'$ . Now here is the whole subject. That $\mathrm{d}v'$ is a perfectly ordinary, entirely Newtonian-looking increment: work it out and you get $\mathrm{d}v' = -u\,\mathrm{d}m/m$ , exactly as Tsiolkovsky would have. There is no relativistic correction to a single burn. Relativity does not make your engine weaker. What relativity does is refuse to let you add up those increments. You made $\mathrm{d}v'$ in the frame you occupied a moment ago; a moment later you occupy a different frame, and you make the next $\mathrm{d}v'$ in that one. To an observer back on Earth, these do not simply pile up — velocity addition is not addition. Compose a boost of $\beta_1$ with a boost of $\beta_2$ and you get $(\beta_1+\beta_2)/(1+\beta_1\beta_2)$ , which is always less than $\beta_1 + \beta_2$ , and always less than 1. But there is a variable in which boosts do just pile up: the rapidity $\varphi = \operatorname{artanh}\beta$ . Rapidity and Hyperbolic Boosts proves the theorem — $\Lambda(\varphi_1)\Lambda(\varphi_2) = \Lambda(\varphi_1 + \varphi_2)$ — and proves that no finite chain of boosts reaches $c$ . A rocket is a finite chain of boosts. That is the whole argument, and the rest is bookkeeping. So the ledger is kept in rapidity. Each infinitesimal burn contributes $\mathrm{d}\varphi = \mathrm{d}v'/c = -(u/c)\,\mathrm{d}m/m$ , these add straightforwardly because rapidities add, and integrating gives $\varphi = (u/c)\ln(m_0/m)$ . The mass ratio buys rapidity, at a rate set by $u/c$ . Then, and only at the very end, you convert to velocity: $\beta = \tanh\varphi$ . And $\tanh$ is bounded. The universe's speed limit is enforced not by an escalating cost, not by mass increasing, not by a force — but by the fact that the function you have to apply at the end has an asymptote.

The core equations

The classical result, for comparison, and to be recovered as a limit:

\Delta v = u\,\ln\frac{m_0}{m} \qquad\text{(Tsiolkovsky, 1903)} \qquad (1)

where: $u$ is the exhaust speed relative to the rocket, $m_0$ the initial mass, $m$ the final mass. Physical reading: unbounded. Nothing in it stops $\Delta v$ from exceeding $c$ ; you need only $\ln(m_0/m) > c/u$ . This is not a small error at high speed — it is a structurally wrong answer, and it is wrong in the direction of promising you something impossible. The relativistic result, in the variable that is actually additive:

\boxed{\;\varphi \;=\; \frac{u}{c}\,\ln\frac{m_0}{m}\;} \qquad (2)

where: $\varphi = \operatorname{artanh}\beta$ is the rapidity. Physical reading: compare (1) and (2) and see that they are the same equation. Tsiolkovsky's $\Delta v/u = \ln(m_0/m)$ has become $\varphi c/u = \ln(m_0/m)$ . Newton's rocket buys $\Delta v$ ; Einstein's buys $\varphi$ . That single substitution is the entire content of the relativistic rocket equation, and everything else follows from what $\varphi$ is. Convert to a velocity — the last step, and the one that installs the speed limit:

\boxed{\;\beta = \tanh\varphi = \tanh\!\left(\frac{u}{c}\,\ln\frac{m_0}{m}\right)\;} \qquad (3)

where: $\tanh$ maps $[0,\infty) \to [0,1)$ . Physical reading: $\beta < 1$ for every finite mass ratio, automatically. No constraint was imposed; the bound is a property of the function. Burn a mass ratio of $10^{100}$ at $u = c$ and you get $\beta = \tanh(230.3) = 1 - 10^{-200}$ : closer to $c$ than any measurement could resolve, and still not $c$ . Invert to get what an engineer actually needs — the fuel bill for a target speed:

\boxed{\;\frac{m_0}{m} = \left(\frac{1+\beta}{1-\beta}\right)^{\!c/2u} = \; k^{\,c/u}, \qquad k \equiv \sqrt{\frac{1+\beta}{1-\beta}} = e^{\varphi}\;} \qquad (4)

where: $k$ is the relativistic Doppler factor for the speed $\beta$ — the same $k$ that reddens a receding ship's light. Physical reading: the mass ratio is the Doppler factor raised to the power $c/u$ , and this single formula contains both the general case and the photon rocket. A photon rocket has $u = c$ , so it pays the Doppler factor once: $m_0/m = k$ . Anything slower pays it $c/u$ times over, and $c/u$ is where the catastrophe lives — for a chemical rocket it is about $67{,}000$ .

Deriving the rocket equation

Derivation AThe burn, in the ship's own frame — and why the increments are rapidities

We want $\varphi = (u/c)\ln(m_0/m)$ . The strategy is the one that makes relativistic mechanics easy: do the physics in the frame where it is trivial, and let rapidity do the composing. The trivial frame is the ship's instantaneous rest frame, where the ship is momentarily at rest, and where — as we will see — the burn looks completely Newtonian.

Step 1. Set up the instantaneous rest frame (IRF). At this instant the rocket has rest mass $m$ and is at rest. It burns for an instant, ejecting exhaust of rest mass $\mu$ at speed $u$ backwards, and ends with rest mass $m + \mathrm{d}m$ (with $\mathrm{d}m < 0$ ) moving forwards at $\mathrm{d}v'$ . Why: the IRF is chosen because "the exhaust speed relative to the rocket" is a statement about this frame — it is how the engine is specified. Any other frame would make $u$ a moving target.

Step 2. Conserve energy in the IRF, to first order:

mc^2 \;=\; (m + \mathrm{d}m)c^2 \;+\; \gamma_u\,\mu c^2 \qquad\Longrightarrow\qquad \mu = -\frac{\mathrm{d}m}{\gamma_u}, \qquad \gamma_u \equiv \frac{1}{\sqrt{1-u^2/c^2}}

Why: the rocket's own kinetic energy after the burn is $\tfrac12 m\,\mathrm{d}v'^2$ , second order in small quantities, and drops out. Note that $\mu \neq -\mathrm{d}m$ : the exhaust's rest mass is smaller than the mass the rocket lost, because part of what the rocket gave up became the exhaust's kinetic energy. Mass is not conserved. Four-momentum is.

Step 3. Conserve momentum in the IRF, to first order:

0 \;=\; m\,\mathrm{d}v' \;-\; \gamma_u\,\mu\, u

Why: the exhaust carries relativistic momentum $\gamma_u \mu u$ backwards, so the rocket must carry $m\,\mathrm{d}v'$ forwards. This is the only place the engine's physics enters.

Step 4. Combine, and watch $\gamma_u$ vanish:

m\,\mathrm{d}v' = \gamma_u\,\mu\,u = \gamma_u \left(-\frac{\mathrm{d}m}{\gamma_u}\right) u = -u\,\mathrm{d}m \qquad\Longrightarrow\qquad \boxed{\;\mathrm{d}v' = -\,u\,\frac{\mathrm{d}m}{m}\;}

Why: this is Tsiolkovsky's differential, exactly, with no relativistic correction whatsoever. The $\gamma_u$ from the exhaust's energy and the $\gamma_u$ from its momentum cancel identically. A single burn is not relativistic. Everything relativistic about a relativistic rocket is in how the burns compose — which is Step 5, and which is the point of the article.

Step 5. Convert the increment to a rapidity increment. The ship's rapidity relative to the launch frame is $\varphi$ ; the burn boosts it by $\mathrm{d}v'$ in its own frame, i.e. by a boost of rapidity $\mathrm{d}\varphi = \operatorname{artanh}(\mathrm{d}v'/c) = \mathrm{d}v'/c$ to first order. And rapidities add under composition of collinear boosts:

\mathrm{d}\varphi \;=\; \frac{\mathrm{d}v'}{c} \;=\; -\,\frac{u}{c}\,\frac{\mathrm{d}m}{m}

Why: this is the step where relativity happens, and it is the only one. We may add $\mathrm{d}\varphi$ to the ship's existing $\varphi$ because $\Lambda(\varphi_1)\Lambda(\varphi_2) = \Lambda(\varphi_1+\varphi_2)$ — the theorem owned by Rapidity and Hyperbolic Boosts. We could not have added $\mathrm{d}v'$ to the ship's existing $v$ , because velocities do not compose that way. Choosing to keep the books in $\varphi$ is not a trick; it is the only variable in which the books close.

Step 6. Integrate. From $m_0$ (at rest, $\varphi = 0$ ) to $m$ :

\int_0^{\varphi}\mathrm{d}\varphi' = -\frac{u}{c}\int_{m_0}^{m}\frac{\mathrm{d}m'}{m'} \qquad\Longrightarrow\qquad \boxed{\;\varphi = \frac{u}{c}\,\ln\frac{m_0}{m}\;}

Why: the integral is the same one Tsiolkovsky did in 1903. Only the thing on the left has changed — and that change is the whole of special relativity, applied to a rocket.

Step 7. Read off the velocity, and the speed limit with it. $\beta = \tanh\varphi$ , so

\beta = \tanh\!\left(\frac{u}{c}\ln\frac{m_0}{m}\right) \;<\; 1 \quad \text{always}.

Why: the bound is free. We did not impose $v < c$ ; we derived a rapidity, and a rapidity's tanh is less than one. Every finite mass ratio gives a finite $\varphi$ , and every finite $\varphi$ gives $\beta < 1$ . This is the same fact as Rapidity and Hyperbolic Boosts' theorem that no finite chain of boosts reaches $c$ — a rocket is such a chain. *Verified in verify_aq-sr-relativistic-rocket-equation.py: the IRF energy and momentum balances are solved symbolically in sympy, confirming $\mathrm{d}v' = -u\,\mathrm{d}m/m$ with $\gamma_u$ cancelling identically; the ODE $\mathrm{d}\varphi/\mathrm{d}m = -(u/c)/m$ is integrated symbolically to $\varphi = (u/c)\ln(m_0/m)$ ; $\beta = \tanh\varphi < 1$ is asserted over 12 decades of mass ratio; and the $c \to \infty$ limit is taken symbolically and shown to give exactly $\Delta v = u\ln(m_0/m)$ .*

Go deeper · Derivation B · 2 minThe fuel bill: the mass ratio is the Doppler factor, raised to the power c over u

Derivation BThe fuel bill: the mass ratio is the Doppler factor, raised to the power c over u

Equation (2) is what a physicist wants; an engineer wants it backwards. Given a target speed, what mass ratio does it cost? Inverting is trivial algebra — but the form it lands in is not trivial at all, and it unifies the general rocket with the photon rocket in a single expression whose meaning is the Doppler shift.

Step 1. Invert equation (2). $\varphi = (u/c)\ln(m_0/m)$ , so

\frac{m_0}{m} = \exp\!\left(\frac{c}{u}\,\varphi\right) = \Big(e^{\varphi}\Big)^{c/u}

Why: the mass ratio is an exponential in the rapidity, and the rate is $c/u$ . That exponent is the entire engineering problem. For a chemical rocket $c/u \approx 67{,}000$ ; for a photon rocket it is 1.

Step 2. Identify $e^{\varphi}$ , which is not just any number. Using $\varphi = \operatorname{artanh}\beta = \tfrac12\ln\frac{1+\beta}{1-\beta}$ :

e^{\varphi} = \sqrt{\frac{1+\beta}{1-\beta}} \;\equiv\; k

Why: $k$ is the relativistic Doppler factor — the factor by which a receding ship's light is reddened, the Bondi $k$ -factor, the exponential of the rapidity. It is one of the most-recurring objects in special relativity, and it has just walked into a rocketry problem uninvited.

Step 3. Assemble the general result:

\boxed{\;\frac{m_0}{m} = k^{\,c/u} = \left(\frac{1+\beta}{1-\beta}\right)^{c/2u}\;}

Why: the mass ratio is the Doppler factor raised to the power $c/u$ . This is one formula, not two, and it says the exhaust speed's only job is to set an exponent. Every factor of $u$ you lose against $c$ is paid for by raising $k$ to a higher power — and $k$ blows up as $\beta \to 1$ , so you are compounding a divergence.

Step 4. Set $u = c$ — the photon rocket, the theoretical best any engine can do, because nothing exits faster than light:

\boxed{\;\frac{m_0}{m} = k = \sqrt{\frac{1+\beta}{1-\beta}}\;}

Why: the photon rocket pays the Doppler factor exactly once, and it is the floor: no engine can do better, because no exhaust can be faster. To reach $\beta = 0.99$ : $k = \sqrt{199} = \mathbf{14.1}$ . Fourteen kilograms of fuel-and-antifuel per kilogram of ship, to reach 99 % of the speed of light. That is not absurd. That is a design.

Step 5. Now spoil it. Take a wildly optimistic fusion drive, $u = 0.1c$ , so $c/u = 10$ and the mass ratio is $k^{10}$ :

\beta = 0.99 \;\Longrightarrow\; \frac{m_0}{m} = (14.107)^{10} = \mathbf{3.12\times10^{11}}

Why: the exponent did that. Ten percent of light speed is a fantastical exhaust — no one knows how to build it — and it still turns a mass ratio of 14 into three hundred billion. The photon rocket is not merely better; it is in a different universe of feasibility, and the reason is that $c/u$ sits in an exponent.

Step 6. And now the chemical rocket, $u = 4.5\ \mathrm{km/s}$ , so $c/u = 66{,}621$ . For a modest $\beta = 0.5$ , $k = \sqrt{3} = 1.732$ :

\frac{m_0}{m} = (1.732)^{66{,}621} = 10^{\,15{,}893}

Why: fifteen thousand, eight hundred and ninety-three orders of magnitude. The observable universe holds around $10^{80}$ atoms. There is no sense in which this is an engineering challenge. Chemical propulsion is not slow for interstellar travel; it is arithmetically excluded, and this one exponent is why. *Verified in verify_aq-sr-relativistic-rocket-equation.py: $m_0/m = k^{c/u}$ is proved symbolically equal to $((1+\beta)/(1-\beta))^{c/2u}$ and to $\exp((c/u)\operatorname{artanh}\beta)$ ; $e^{\varphi} = k$ is confirmed to machine precision; and the printed mass ratios (14.107, 3.12e11, 10^15893, 243, 4.359, 1.732) are each recomputed and asserted against the literals above.*

Figure

THE EXPONENT IS THE WHOLE STORY. Mass ratio needed to reach a given speed, on a log scale that spans 16,000 decades. The mass ratio is the Doppler factor k raised to the power c/u — so the exhaust speed does nothing but set an EXPONENT, and every factor you lose against c is compounded. Photon rocket (u = c, the theoretical floor): reaching 0.99c costs a mass ratio of just 14.1 — a design, not a fantasy. A fantastical fusion drive (u = 0.1c) costs 3.1e11 for the same speed. A chemical rocket (u = 4.5 km/s, c/u = 66,621) needs 10 to the power 15,893 to reach even HALF light speed — and the observable universe contains about 10 to the 80 atoms. Chemical propulsion is not slow for interstellar flight; it is arithmetically excluded

Limiting cases

Case	Result	Meaning
$\varphi \ll 1$ ( $c \to \infty$ )	$\Delta v \to u\ln(m_0/m)$	Tsiolkovsky, 1903, recovered exactly
any finite $m_0/m$	$\beta = \tanh\varphi < 1$	the speed limit, for free — no fuel supply reaches $c$
$m_0/m \to \infty$	$\beta \to 1^-$ , never $=1$	you approach $c$ logarithmically slowly and never arrive
photon rocket, $u = c$	$m_0/m = k = \sqrt{\tfrac{1+\beta}{1-\beta}}$	the floor: no exhaust is faster than light
$u = c$ , $\beta = 0.99$	$m_0/m = 14.107$	astonishingly cheap — and the best physics allows
$u = 0.1c$ , $\beta = 0.99$	$m_0/m = 3.12\times10^{11}$	the exponent $c/u = 10$ does this
$u = 4.5$ km/s, $\beta = 0.5$	$m_0/m = 10^{15{,}893}$	chemical propulsion, arithmetically excluded
$u \to 0$	$m_0/m \to \infty$ for any $\beta > 0$	a rocket with no exhaust speed goes nowhere
single burn, IRF	$\mathrm{d}v' = -u\,\mathrm{d}m/m$	exactly Newtonian — no relativistic correction

Common trapVelocities do not add — so do not add them. This is the error the equation exists to prevent

The seductive wrong move is to take Tsiolkovsky's $\Delta v = u\ln(m_0/m)$ , notice that it can exceed $c$ , and then patch it — divide by a $\gamma$ somewhere, or cap it, or declare the rocket's mass to grow. All of these are wrong, and the last one is wrong twice. Here is what is actually true, and it is stranger and simpler than any patch. A single burn is exactly Newtonian. In the ship's instantaneous rest frame, $\mathrm{d}v' = -u\,\mathrm{d}m/m$ , with the $\gamma_u$ factors from the exhaust's energy and momentum cancelling identically (Step 4). The engine is not weakened by relativity, at any speed. Relativity never touches the burn. It touches the sum. Because the ship changes frames between burns, the increments cannot be added as velocities. They must be added in the variable that composes additively under boosts, and that variable is the rapidity $\varphi = \operatorname{artanh}\beta$ . So $\ln(m_0/m)$ buys $\varphi$ , not $v$ — and $v = c\tanh\varphi$ is bounded because $\tanh$ is bounded. Two corollaries worth stating flatly:

The rocket's mass does not increase with speed. There is no relativistic mass here or anywhere; $m$ is the invariant rest mass, and it decreases, because the ship is throwing it out of the back. (Relativistic Force owns the argument that kills relativistic mass for good.)
Mass is not conserved in the burn. The exhaust's rest mass $\mu = -\mathrm{d}m/\gamma_u$ is less than the mass the rocket lost; the difference became the exhaust's kinetic energy. Four-momentum is conserved. Rest mass is not, and never was.

Worked examples

1 · The photon rocket to Proxima Centauri, at 99 % of light speed

What mass ratio does an ideal photon rocket need to reach $\beta = 0.99$ ?

Step 1 — identify. $u = c$ (the theoretical optimum), so $c/u = 1$ and equation (4) collapses to $m_0/m = k$ .

Step 2 — apply. $k = \sqrt{\dfrac{1+0.99}{1-0.99}} = \sqrt{\dfrac{1.99}{0.01}} = \sqrt{199} = \mathbf{14.107}$ .

Step 3 — read it. For every kilogram of ship arriving at $0.99c$ , you must have started with $14.1\ \mathrm{kg}$ — that is $13.1\ \mathrm{kg}$ of fuel converted entirely into a beam of light aimed backwards. The rapidity bought is $\varphi = \operatorname{artanh}(0.99) = 2.6467$ , and indeed $e^{2.6467} = 14.107$ . ✓

Step 4 — the boundary number. Compare the Saturn V: $3{,}038{,}500\ \mathrm{kg}$ on the pad, $118{,}000\ \mathrm{kg}$ to low Earth orbit — a mass ratio of about 26, larger than 14.1, to reach a few kilometres per second. The photon rocket reaches 99 % of the speed of light on a smaller mass ratio than the Saturn V needed to reach orbit. The difficulty of interstellar flight is not the mass ratio. It is that we have no idea how to make an engine whose exhaust is light and whose fuel is half antimatter. The kinematics is easy; the engineering is the impossible part — and it is worth knowing which is which.

2 · What Tsiolkovsky would have promised, and what he would have got

A chemical rocket, $u = 4.5\ \mathrm{km/s}$ , burns a mass ratio of $m_0/m = 20$ — an ambitious but real launch vehicle. Compare the Newtonian and relativistic answers.

Step 1 — Newtonian. Equation (1): $\Delta v = 4.5 \times \ln 20 = 4.5 \times 2.996 = \mathbf{13.48\ km/s}$ .

Step 2 — relativistic. Equation (2): $\varphi = (u/c)\ln 20 = (4.5/299{,}792.458)\times 2.996 = 4.497\times10^{-5}$ . Then $\beta = \tanh\varphi = 4.497\times10^{-5}$ to eleven significant figures, giving $v = \mathbf{13.48\ km/s}$ .

Step 3 — the boundary number. They agree to about one part in $10^{9}$ . The relativistic correction is $\varphi^3/3 \approx 3\times10^{-14}$ , which is $10^{-9}\ \mathrm{m/s}$ — utterly irrelevant, and that is the correct answer. Relativity does not make rocketry harder at rocket speeds. It changes nothing you can measure until $\varphi \sim 1$ , i.e. until $\ln(m_0/m) \sim c/u \approx 66{,}621$ , i.e. until the mass ratio is around $e^{66621}$ . Newton is not an approximation that is slightly wrong here — he is right to nine decimal places, and the relativistic equation's entire job is to be right in a regime no chemical rocket will ever enter.

Figure

THE LADDER THAT NEVER REACHES THE TOP. Left: each burn adds the SAME rapidity increment — rapidity is what the rocket buys, and it grows without bound, linearly in ln(m₀/m). Right: the velocity you actually get is β = tanh φ, and tanh has an asymptote at 1. So equal rungs on the rapidity ladder map onto rungs that crowd ever closer to c, and never touch it. Ten boosts of rapidity 0.1 do not give 1.0c — they give tanh(1.0) = 0.762c. The speed limit is not enforced by a force, or by mass increasing (it does not), or by the engine weakening (it does not). It is enforced by the shape of a hyperbolic tangent

Physics of the portrait

The equation is 1903. The man who wrote it was a deaf, self-taught schoolteacher in a provincial Russian town, and he had the whole of spaceflight essentially right.

Origins · 2 minKonstantin Tsiolkovsky and how this was discovered

Konstantin Tsiolkovsky

1857–1935 · the rocket equation · staging · and a schoolteacher who worked out spaceflight before anyone had flown

Konstantin Eduardovich Tsiolkovsky lost most of his hearing to scarlet fever at ten, was refused schooling because of it, and educated himself in the libraries of Moscow. He spent his working life teaching mathematics in Kaluga, a hundred miles from the capital, and in his spare time he worked out the theory of spaceflight. In May 1903, in the journal Nauchnoe Obozrenie (Science Review), he published "Exploration of Outer Space by Means of Rocket Devices" — the paper that contains equation (1). It preceded Esnault-Pelterie (1913) and Goddard (1919), and it appeared in the same year the Wright brothers first left the ground at Kitty Hawk: the mathematics of getting to orbit was published before anyone had flown an aeroplane. In it Tsiolkovsky not only derived $\Delta v = u\ln(m_0/m)$ but drew the correct conclusion from it — that a single-stage rocket cannot reach orbit on chemical fuel, and that the answer is staging, throwing away the empty tanks. He computed the $8\ \mathrm{km/s}$ needed for orbit and specified liquid oxygen and liquid hydrogen as the propellant. Every one of those calls was right. What he could not have known is that his equation is the $c \to \infty$ limit of a deeper one, in which the logarithm of the mass ratio buys not velocity but rapidity — a quantity that had been defined only in 1911, and which would not be pointed at a rocket until Ackeret did it in 1946. Tsiolkovsky's equation survives inside the relativistic one, unchanged, with a single symbol swapped.

History

1903

— In May, Tsiolkovsky publishes "Exploration of Outer Space by Means of Rocket Devices" in Nauchnoe Obozrenie, giving the rocket equation and arguing for staging and liquid hydrogen and oxygen.

1903

— In December, the Wright brothers fly at Kitty Hawk; the theory of spaceflight is older than powered flight.

1905

— Einstein's relativity makes velocity addition non-linear, and therefore makes Tsiolkovsky's equation an approximation.

1911

— Robb introduces rapidity as the additive measure of a boost, the variable the relativistic rocket equation is written in.

1913

— Esnault-Pelterie publishes independently on rocket flight; Goddard follows in 1919.

1946

— Ackeret gives the general derivation of the relativistic rocket equation in Helvetica Physica Acta, volume 19, page 103.

1969

— The Saturn V reaches the Moon on a launch-to-orbit mass ratio of order 25, a purely Newtonian machine operating nine orders of magnitude below any relativistic correction.

Going further

Texts. Rindler, Introduction to Special Relativity, 2nd ed., §31 · Taylor & Wheeler, Spacetime Physics, 2nd ed., ch. 8 · J. Ackeret, "Zur Theorie der Raketen", Helvetica Physica Acta 19, 103 (1946) · K. E. Tsiolkovsky, "Exploration of Outer Space by Means of Rocket Devices", Nauchnoe Obozrenie 5 (1903).

Provenance. Signature $(+,-,-,-)$ ; $P\cdot P = +m^2c^2$ ; $m$ is the invariant rest mass, which decreases only because the ship ejects it. No relativistic mass anywhere. The AllQFT base was re-derived in full and found correct — all seven of its equation cards check out. verify_aq-sr-relativistic-rocket-equation.py proves $\mathrm{d}v' = -u\,\mathrm{d}m/m$ symbolically (with $\gamma_u$ cancelling), integrates the rapidity ODE to $\varphi = (u/c)\ln(m_0/m)$ , proves $m_0/m = k^{c/u} = ((1+\beta)/(1-\beta))^{c/2u}$ and $e^\varphi = k$ , takes the $c\to\infty$ limit symbolically to recover Tsiolkovsky, and recomputes every printed number: $14.107$ , $3.12\times10^{11}$ , $10^{15{,}893}$ , $243$ , $\sqrt3 = 1.732$ , $\varphi(0.99) = 2.6467$ , $c/u = 66{,}621$ , and the worked example's $13.48\ \mathrm{km/s}$ agreeing Newtonian-to-relativistic to one part in $10^9$ .

Scope. This article derives the rocket equation and the mass ratio. It does not treat the time the journey takes, proper acceleration profiles, or the twin-paradox bookkeeping of a round trip — a constant-proper-acceleration ship is the Rindler hyperbola, and the S6 article on relativistic uniform acceleration owns it. Engine physics, radiation shielding and the antimatter storage problem are engineering, and are out of scope for all of physics.

Builds on. Relativistic Four-Momentum · Relativistic Momentum · Relativistic Energy · Energy-Momentum Relation · Mass-Energy Equivalence · Rapidity and Hyperbolic Boosts. Leads to. Four-Acceleration · Relativistic Force · Massless Particles.

Does the rocket get heavier as it speeds up, making it harder to push?

No. That is the relativistic-mass story and it is wrong — Relativistic Force owns the argument that kills it. The rocket's mass $m$ in these equations is its invariant rest mass, the same quantity in every frame, and it decreases over the burn, for the wholly mundane reason that the ship is throwing pieces of itself out of the back. Nor does the engine weaken: Step 4 of the derivation shows that a single burn, in the ship's own instantaneous rest frame, produces exactly the Newtonian $\mathrm{d}v' = -u\,\mathrm{d}m/m$ , with the $\gamma_u$ factors cancelling identically. Relativity does not make the rocket harder to push. It makes the pushes harder to add up.

Where exactly does relativity enter, if a single burn is Newtonian?

In precisely one place: the composition of the burns. Each increment $\mathrm{d}v'$ is made in a different frame — the frame the ship occupied at that instant — and to get the ship's speed in the launch frame you must compose all those boosts. Velocities do not compose additively; rapidities do. So the correct ledger is kept in $\varphi = \operatorname{artanh}\beta$ , and $\ln(m_0/m)$ buys $\varphi$ rather than $v$ . Everything else in the article — the speed limit, the $\tanh$ , the brutal exponent $c/u$ — is a consequence of that one substitution. It is a remarkably clean example of relativity changing not the dynamics but the arithmetic.

Is the exhaust's rest mass equal to the mass the rocket lost?

No, and this catches people. The exhaust's rest mass is $\mu = -\mathrm{d}m/\gamma_u$ , which is smaller than the rocket's mass loss $|\mathrm{d}m|$ by the factor $\gamma_u$ . The difference went into the exhaust's kinetic energy. Rest mass is simply not a conserved quantity — four-momentum is — and this is the same lesson as Binding Energy and Mass Defect: a system's invariant mass is not the sum of its parts' masses, and mass and energy are not two separate ledgers to be balanced against each other. For a photon rocket, taken to the limit, $\gamma_u \to \infty$ and $\mu \to 0$ : the exhaust has no rest mass at all, and the entire mass loss became light.

Why is the photon rocket the best possible, and how can it be so cheap?

It is the best possible because the exponent in $m_0/m = k^{c/u}$ is minimised when $u$ is maximised, and no exhaust can exceed $c$ . So $u = c$ gives the floor, $m_0/m = k$ , and there is no engine anywhere in physics that beats it. As for cheap: $k = \sqrt{(1+\beta)/(1-\beta)} = 14.1$ at $\beta = 0.99$ — a smaller mass ratio than the Saturn V's ≈ 26. The kinematics of relativistic flight is genuinely easy; it is the engine that is impossible. You would need to annihilate matter with antimatter and collimate the resulting gamma rays into a perfect backward beam, at essentially 100 % efficiency, and store the antimatter in the meantime. Every one of those is beyond us by many orders of magnitude. But the mass ratio was never the obstacle, and it is worth being clear about which part of the problem is hard.

Why does the mass ratio blow up so violently for a slow exhaust?

Because $c/u$ is an exponent, not a coefficient. $m_0/m = k^{c/u}$ , so halving your exhaust velocity does not double the fuel bill — it squares it. A fusion drive at $u = 0.1c$ has $c/u = 10$ , turning a photon rocket's $14.1$ into $14.1^{10} = 3.1\times10^{11}$ . A chemical rocket has $c/u \approx 66{,}621$ , and raising even the modest $k = \sqrt3$ (for $\beta = 0.5$ ) to that power gives $10^{15{,}893}$ . This is why exhaust velocity, and nothing else, is the figure of merit for interstellar propulsion — and why the only alternatives that work are the ones that carry no reaction mass at all, like a beam-pushed light sail.

Can I avoid the rocket equation by scooping up fuel along the way?

That is the Bussard ramjet, and the rocket equation does not forbid it — you have changed the problem, because $m_0/m$ is no longer fixed by what you carry. What kills it in practice is drag: to scoop interstellar hydrogen you must interact with it, and the momentum you take from the medium in order to collect it turns out (in the usual analyses) to cost more than the fusion energy you extract from it, so the "ramjet" works better as a brake than as an engine. But note the structure of the escape: the ramjet does not beat the rocket equation, it sidesteps it, and so does a light sail, and so does a laser-pushed sail. Every serious relativistic-flight proposal is an attempt not to be a rocket — which is, in its way, the strongest possible testimonial to how severe that exponent is.