Mathematics › 🎓 University Year 1 › Improper Integrals: When Infinity Still Adds Up to a Finite Number

Improper Integrals: When Infinity Still Adds Up to a Finite Number

Stretch the region under a curve out to infinity — sometimes the area is unbounded, and sometimes it settles on a finite total.

University Year 1Calculus II / Linear Algebra

Improper Integrals: When Infinity Still Adds Up to a Finite Number — illustration

💡

The big idea: An ordinary definite integral runs between two finite numbers. An improper integral pushes one of those bounds to infinity (or lets the integrand blow up), and asks a subtle question: does the accumulating area approach a finite limit, or grow without bound? The answer is decided by a limit, and for the family 1/xp it hinges on a single knife-edge at p = 1.

🎯 By the end, you'll be able to

Define an improper integral as the limit of ordinary integrals
Decide convergence or divergence by evaluating that limit
Apply the p-test to integrals of 1/xᵖ
Connect improper integrals to the convergence of series like Σ 1/nᵖ

📎 You should already know

The definite integral and antiderivatives
Limits at infinity
Series and partial sums

An area that never ends

Consider the region under a curve that trails off to the right forever — the bound is not a number but ∞. Its area is an infinitely long strip, yet the strip gets thinner and thinner. Does its total area explode, or does it converge on something finite?

Astonishingly, an infinitely long region can enclose a finite area, provided it thins out fast enough. Deciding which way it goes is what improper integrals are about.

🔑 An improper integral is a limit

An integral to infinity is defined as the limit of ordinary integrals with a finite upper bound b, as b marches off to infinity. If that limit is a finite number, the integral converges; if the limit is infinite or does not exist, it diverges.

\[ \int_{a}^{\infty} f(x)\,dx=\lim_{b\to\infty}\int_{a}^{b} f(x)\,dx \]

Compute the ordinary integral up to b, then let b → ∞. The value of that limit decides convergence.

🎮 Improper Limits LIVE

Push a bound to infinity and watch partial sums converge — or diverge.

The tale of two curves

The cleanest way to feel this is to compare 1/x² and 1/x from 1 to infinity. They look similar — both decay to zero — but one converges and the other does not. The difference is how fast they decay. This single comparison is the heart of the whole topic.

📝 Worked example: Does ∫₁^∞ 1/x² dx converge, and if so to what?

Integrate to a finite bound b: \( \int_{1}^{b} x^{-2}\,dx=\Big[-\tfrac{1}{x}\Big]_{1}^{b}=1-\tfrac{1}{b} \).
Now take the limit: \( \lim_{b\to\infty}\big(1-\tfrac{1}{b}\big)=1-0=1 \).
The limit is finite, so the integral converges.

✓ It converges to 1 — an infinitely long region with a finite area of exactly 1.

📝 Worked example: Does ∫₁^∞ 1/x dx converge?

Integrate to b: \( \int_{1}^{b}\tfrac{1}{x}\,dx=\big[\ln x\big]_{1}^{b}=\ln b-\ln 1=\ln b \).
Take the limit: \( \lim_{b\to\infty}\ln b=\infty \).
The limit is infinite, so the area grows without bound.

✓ It diverges — even though 1/x → 0, it does so too slowly for the area to be finite.

✨ The p-test: everything turns at p = 1

For the family 1/x^p, the integral ∫₁^∞ x^−p dx converges exactly when p > 1 and diverges when p ≤ 1. So 1/x² (p = 2) converges but 1/x (p = 1) diverges — the boundary sits precisely at p = 1. Curiously, over 0 to 1 the rule flips: ∫₀¹ x^−p dx converges when p < 1 instead.

\[ \int_{1}^{\infty}\frac{1}{x^{p}}\,dx \;\text{ converges} \iff p>1 \]

The p-test for improper integrals at infinity: convergence requires decay faster than 1/x.

The same knife-edge governs series

The parallel with infinite sums is exact. The p-series Σ 1/n^p converges precisely when p > 1 — the same threshold. In particular the harmonic series Σ 1/n (p = 1) diverges, mirroring ∫ 1/x. A convergent example is the geometric series: adding 1 + ½ + ¼ + … gives partial sums creeping toward a finite ceiling.

📝 Worked example: What does the geometric series 1 + ½ + ¼ + ⅛ + … add up to?

This is \( \sum_{n=0}^{\infty} r^{n} \) with ratio \( r=\tfrac12 \), and \( |r|<1 \) so it converges.
The sum of a convergent geometric series is \( \dfrac{1}{1-r} \).
Substitute \( r=\tfrac12 \): \( \dfrac{1}{1-\tfrac12}=\dfrac{1}{\tfrac12}=2 \).

✓ The partial sums converge to 2 — they approach it but never exceed it.

⚠️ Terms going to zero is not enough

A tempting mistake: “the terms shrink to 0, so the sum must be finite.” The harmonic series Σ 1/n is the counterexample — its terms go to 0 yet the sum is infinite. Shrinking terms are necessary for convergence but not sufficient; what matters is how fast they shrink.

Check your understanding

1. How is the improper integral ∫ₐ^∞ f(x) dx defined?

You integrate to a finite bound b and then take the limit b → ∞; a finite limit means convergence.

2. What is the value of ∫₁^∞ 1/x² dx?

∫₁ᵇ x⁻² dx = 1 − 1/b, and as b → ∞ this approaches 1. The integral converges to 1.

3. Does ∫₁^∞ 1/x dx converge or diverge?

∫₁ᵇ 1/x dx = ln b, which grows to ∞. 1/x decays too slowly, so the integral diverges.

4. For which p does ∫₁^∞ 1/xᵖ dx converge?

The p-test: at infinity the integral converges exactly when p > 1 (decay faster than 1/x) and diverges when p ≤ 1.

5. The geometric series 1 + ½ + ¼ + ⅛ + … converges to:

With first term 1 and ratio r = ½, the sum is 1/(1 − ½) = 2. The partial sums approach 2 without reaching it.

✅ Key takeaways

An improper integral to infinity is the limit of finite integrals; a finite limit means it converges.
∫₁^∞ 1/x² dx = 1 (converges) but ∫₁^∞ 1/x dx = ∞ (diverges) — decay speed decides everything.
The p-test: ∫₁^∞ 1/xᵖ dx converges exactly when p > 1.
The same threshold governs p-series Σ 1/nᵖ; the harmonic series Σ 1/n (p = 1) diverges.
Terms shrinking to zero is necessary but not sufficient for a sum to converge — how fast they shrink is what counts.