Improper Integrals - Convergence

Core Idea

An improper integral has either an infinite limit of integration or an integrand with an infinite discontinuity in the interval. It is evaluated as a limit: the integral from a to infinity of f(x) dx = lim(b->infinity) of the integral from a to b of f(x) dx. If this limit exists and is finite, the integral converges; otherwise, it diverges. The p-integral (integral of 1/x^p from 1 to infinity) converges if and only if p > 1, a key benchmark.

How It's Best Learned

Start with concrete examples: integral of 1/x² from 1 to infinity (converges to 1) vs. integral of 1/x from 1 to infinity (diverges). Evaluate by antidifferentiating and taking the limit. Practice both types of impropriety (infinite bounds and discontinuous integrands). Introduce the p-test as a reference point.

Common Misconceptions

• Evaluating an improper integral without taking a limit (plugging in infinity directly).
• Not recognizing an infinite discontinuity within the interval (e.g., integral of 1/x from -1 to 1 has a discontinuity at 0).
• Confusing convergence of the integral with convergence of the integrand to zero (the integrand can go to zero and the integral still diverge).

Explainer

The Fundamental Theorem of Calculus tells you how to evaluate ∫ₐᵇ f(x) dx: find an antiderivative and plug in the limits. But this recipe assumes f is continuous on a closed, bounded interval [a, b]. An improper integral violates at least one of those conditions — either a limit is ±∞, or the integrand blows up somewhere in the interval. Because you can't "plug in" infinity, you replace the problematic boundary with a parameter and take a limit.

For an infinite upper limit: ∫₁^∞ f(x) dx = lim_{b→∞} ∫₁^b f(x) dx. If the limit exists and is finite, the integral converges to that value; otherwise it diverges. The p-integral ∫₁^∞ 1/xᵖ dx is the benchmark. When p > 1, the antiderivative is x^1−p/(1−p), which goes to 0 as x → ∞, giving a finite answer: it converges to 1/(p−1). When p = 1, the antiderivative is ln(x), which grows without bound — diverges. When p < 1, even worse divergence. So the rule is: ∫₁^∞ 1/xᵖ dx converges if and only if p > 1.

The most common error is forgetting that the integrand going to zero is necessary but not sufficient for convergence. The function 1/x → 0 as x → ∞, yet ∫₁^∞ 1/x dx diverges. Intuitively, 1/x shrinks, but it shrinks too slowly — the accumulation outpaces the decay. In contrast, 1/x² shrinks fast enough that the infinite tail has finite total area. The distinction between "slow decay" and "fast decay" is precisely what the p-test captures.

The second type of improper integral involves an integrand with an infinite discontinuity inside the interval. Consider ∫₀¹ 1/√x dx: the integrand blows up at x = 0. Replace the problem boundary with a parameter: lim_{a→0⁺} ∫ₐ¹ 1/√x dx = lim_{a→0⁺} [2√x]ₐ¹ = 2 − 0 = 2, which converges. The key habit is always checking: does the integrand have any discontinuities on the interval, including at the endpoints? A subtle discontinuity buried inside an interval — like 1/x on [−1, 1] — is easy to miss, but naively applying the FTC gives the wrong answer of 0 (when the integral actually diverges).