Case Study 2 — Escape Velocity: Doing Work All the Way to Infinity
Field: Physics (gravitation and orbital mechanics) Calculus used: Type 1 improper integrals (Section 17.1), the p-integral rule at infinity (Section 17.1), the work integral (preview of Chapter 18), comparison reasoning (Section 17.3)
A mission engineer is asked a deceptively simple question: how fast must a spacecraft be moving when it leaves the launch pad so that it never falls back to Earth — not after an hour, not after a century, not ever? "Never" is the word that makes this a calculus problem rather than an algebra problem. To escape forever, the craft must climb against gravity over an infinite distance, and the total work done against an ever-weakening force is an improper integral. Whether that integral converges decides whether escape is possible at all.
The force that never quite quits
Newton's law of gravitation says that a body of mass $m$ at distance $r$ from the center of a planet of mass $M$ feels an inward pull
$$F(r) = \frac{GMm}{r^2},$$
where $G$ is the gravitational constant. The crucial feature is the $1/r^2$: the force weakens as the craft climbs, but it never reaches zero. At ten Earth radii it is a hundredth as strong; at a thousand radii, a millionth — yet always present, always pulling back. To leave for good, the spacecraft must do work against this pull at every point of an infinitely long journey. The total work is the integral of force over distance, taken all the way out:
$$W = \int_{r_0}^\infty \frac{GMm}{r^2}\,dr,$$
where $r_0$ is the launch radius (for a surface launch, the planet's radius). This is an honest Type 1 improper integral in the sense of Section 17.1: the interval $[r_0, \infty)$ is unbounded. The whole question of escape hinges on whether sweeping the upper limit out to infinity yields a finite total — because a finite amount of work can be supplied by a finite launch speed, while an infinite amount cannot be supplied at all.
Evaluating the work
Pull the upper limit back to a finite wall at $r = t$, evaluate the proper integral, then take the limit — the only honest procedure (Section 17.1):
$$\int_{r_0}^t \frac{GMm}{r^2}\,dr = GMm\left[-\frac{1}{r}\right]_{r_0}^t = GMm\left(\frac{1}{r_0} - \frac{1}{t}\right).$$
Now slide the wall outward. As $t \to \infty$, the term $1/t \to 0$, and the accumulated work approaches a finite ceiling:
$$W = \lim_{t\to\infty} GMm\left(\frac{1}{r_0} - \frac{1}{t}\right) = \frac{GMm}{r_0}.$$
The integral converges, and the reason is exactly the p-integral rule of Section 17.1: the integrand decays like $1/r^2$, a power with $p = 2 > 1$, comfortably on the convergent side of the borderline. The region under the force curve from $r_0$ to infinity is infinitely long but has finite area $GMm/r_0$. This finite number is the total energy the craft must spend to break free — equivalently, the magnitude of the gravitational potential energy $U(r_0) = -GMm/r_0$ that binds it.
From work to escape velocity
Energy conservation finishes the job. If the craft leaves the surface with speed $v_e$, its kinetic energy $\tfrac12 m v_e^2$ must be at least the work required to reach infinity (where, in the marginal case, it arrives with zero speed to spare). Setting kinetic energy equal to the escape work:
$$\frac{1}{2}m v_e^2 = \frac{GMm}{r_0} \;\Longrightarrow\; v_e = \sqrt{\frac{2GM}{r_0}}.$$
The mass $m$ cancels — a feather and a freighter need the same escape speed, the gravitational echo of Galileo's falling bodies. Using the surface-gravity shortcut $g = GM/r_0^2$, we can rewrite this as $v_e = \sqrt{2 g r_0}$. For Earth, with $g \approx 9.81\ \text{m/s}^2$ and $r_0 \approx 6.371\times 10^6\ \text{m}$:
$$v_e = \sqrt{2 \cdot 9.81 \cdot 6.371\times 10^6} \approx \sqrt{1.250\times 10^8} \approx 1.118\times 10^4\ \text{m/s} \approx 11.2\ \text{km/s}.$$
About eleven kilometers per second — roughly Mach 33. That single number, the gateway to interplanetary travel, is the value of a convergent improper integral.
The borderline that would have changed everything
Here is where the abstract convergence threshold of Section 17.1 acquires literal, physical teeth. Suppose, counterfactually, that gravity fell off only like $1/r$ instead of $1/r^2$ — that the force law were $F(r) = k/r$. Then the escape work would be
$$W = \int_{r_0}^\infty \frac{k}{r}\,dr = k\lim_{t\to\infty}\big[\ln r\big]_{r_0}^t = k\lim_{t\to\infty}(\ln t - \ln r_0) = \infty.$$
This is precisely the divergent p-integral with $p = 1$ — the logarithm that grows without bound, the same razor's-edge divergence as $\int_1^\infty \frac{1}{x}\,dx$ in Worked Example 17.1.2. In such a universe the work to escape would be infinite, and no finite launch speed could ever break free. Every object would be gravitationally bound forever, no matter how fast it started. Galaxies, rockets, light itself — all trapped. We live in an escapable universe for one mathematical reason: the gravitational exponent sits at $p = 2$, on the convergent side of the line $p = 1$ that this entire chapter circles. The boundary between $p > 1$ and $p \le 1$ is not a textbook abstraction; in this application it is the boundary between a cosmos you can leave and one you cannot.
Where the comparison test earns its keep
Real spacecraft do not climb through empty space; at low altitude they push through atmosphere, and a fuller model adds a drag term to the force. Suppose the engineer's force model is some messy $F(r)$ with no elementary antiderivative, but she can show that for large $r$ it is bounded above by $C/r^2$ for some constant $C$. Then by direct comparison (Section 17.3),
$$\int_{r_0}^\infty F(r)\,dr \;\le\; \int_{r_0}^\infty \frac{C}{r^2}\,dr = \frac{C}{r_0} < \infty,$$
so the escape work is finite even though she never found a formula for it. The comparison test certifies that escape remains possible — the only question that ultimately matters — without demanding an exact value. This is the working physicist's habit: when the integral resists evaluation, bound it against a p-integral you understand and read off convergence. The same logic that tells a statistician a Gaussian tail is finite (Case Study 1) tells the engineer that her thrust budget is finite.
A parallel in a far-off field
The structure here — a quantity accumulated over infinite time or distance, finite only because of fast enough decay — is not unique to gravity. In economics, the present value of a perpetuity that pays at continuous rate $R$ forever, discounted at rate $r > 0$, is the improper integral $\int_0^\infty R e^{-rt}\,dt = R/r$. It converges for the same reason the escape work does: the integrand decays fast enough (exponentially here, like $1/r^2$ there) that an infinite horizon still yields a finite total. And just as a $1/r$ gravity law would make escape impossible, a zero discount rate ($r = 0$) makes the perpetuity's value infinite — the integral diverges, and "money forever" becomes literally priceless. The mathematics of "work to infinity" and "value to infinity" is one mathematics, wearing different physical clothes.
Why this matters
Escape velocity is the founding calculation of spaceflight, and it is improper integration in its purest applied form: a force that never vanishes, integrated over a distance that never ends, yielding a finite answer because the decay rate clears the convergence bar. The lesson generalizes. Gravitational binding energy, the self-energy of a charge, the total radiation from a source, the lifetime exposure to a decaying drug — each is a quantity spread over an unbounded domain, each converges or diverges by the rate-of-decay verdict of Section 17.1, and each can be certified finite by comparison even when it cannot be evaluated. The p-integral threshold $p = 1$ is, in this chapter's most vivid application, the dividing line between freedom and capture.
Discussion Questions
- Recompute $v_e$ for the Moon ($g \approx 1.62\ \text{m/s}^2$, $r_0 \approx 1.74\times 10^6\ \text{m}$). Why is it so much smaller than Earth's, and what does that imply for launching from the Moon?
- The work integral gave $W = GMm/r_0$, which is also $|U(r_0)|$, the depth of the gravitational potential well. Explain why "work to escape" and "binding energy" are the same improper integral viewed from two directions.
- For a force law $F(r) = k/r^p$, find every value of $p$ for which escape requires only finite work. Connect your answer directly to the p-integral rule of Section 17.1.
- A black hole's escape velocity reaches the speed of light at the event horizon. Without relativity, what does the Newtonian formula $v_e = \sqrt{2GM/r_0}$ predict for the radius at which $v_e = c$? (This recovers the Schwarzschild radius up to a constant — a famous coincidence.)
- Suppose atmospheric drag makes the true force larger than $C/r^2$ near the surface but the model is only valid out to some finite altitude. Does the divergence-or-convergence verdict depend on near-surface behavior, or only on the tail? Justify using how improper integrals are decided.
Short Annotated Reading
- Halliday, Resnick, & Walker (2014). Fundamentals of Physics, ch. on gravitation. Derives escape velocity and gravitational potential energy as the work integral done here; the cleanest physics-side companion.
- Taylor, J. R. (2005). Classical Mechanics, ch. 4 (energy). Treats conservative forces and potential energy as integrals to infinity, making the improper-integral structure explicit.
- Stewart, J. (2020). Calculus: Early Transcendentals (9th ed.), §6.4 (Work) and §7.8 (Improper Integrals). The calculus scaffolding: work as an integral, then the infinite-interval limit.