Every definite integral so far had finite limits and a continuous integrand on a closed interval. An improper integral breaks one of those rules: a limit of integration is infinite, as in , or the integrand is unbounded on the interval, as in . Calculus handles both the same way: integrate on a safe interval, then take a limit.
Infinite limits of integration
If the limit is a finite number, the improper integral converges to that number. If the limit is infinite or does not exist, the integral diverges. The same idea works at the other end: .
Worked example 1. Evaluate or show that it diverges.
The limit is a finite number, so the integral converges to 1.
Worked example 2. Evaluate or show that it diverges.
, so the integral diverges.
The figure puts the two side by side. Both graphs approach 0, yet one area levels off and the other grows without bound. A height that goes to 0 is not enough to make the area finite: only the limit of the integrals decides.
Figure 1. The area under 1/x² from 1 to b approaches 1; the area under 1/x is ln b, which grows without bound. LevlPrep original diagram.
Powers of x: the p-integrals
For , . When , and the integral converges to . When , and it diverges; gives , which also diverges.
Integral
equals
As
converges to 1
converges to 2
diverges
diverges
(from 0)
converges to 1
Integrands that blow up
If is unbounded at an endpoint, put the limit at that endpoint. For an upper endpoint : .
Worked example 3. Evaluate .
The integrand is unbounded as . With :
The integral converges.
The dangerous case is a blow-up inside the interval. looks like , a negative answer for a positive integrand. The Fundamental Theorem does not apply, because is not continuous at 0. Look at the piece from 0 to 1: , so the integral diverges. Before integrating, always check the integrand for vertical asymptotes on the interval.