When you first learn integration, the rules feel reassuringly clean: you integrate a continuous function over a closed, bounded interval and you get a number. But mathematics — and the real world — rarely stays that tidy. What happens when the interval of integration stretches out to infinity? Or when the function you are integrating blows up to infinity somewhere inside the interval? These situations are exactly where improper integrals come in, and mastering them is a genuine milestone in calculus. From computing the area under probability density functions in statistics to evaluating the Laplace transform in engineering, improper integrals appear wherever a process runs over an unbounded domain.
An improper integral is any definite integral that either has one or both limits of integration equal to infinity, or whose integrand (the function being integrated) becomes infinite at one or more points inside the interval of integration. In both cases, the standard Riemann integral — as you learned it for proper integrals — no longer applies directly. The key idea is to replace the problematic boundary with a variable and then take a limit.
By the end of this guide you will be able to identify whether an integral is improper, evaluate it using limits, and determine whether it converges (gives a finite answer) or diverges (does not). This page covers both Type I (infinite limits) and Type II (discontinuous integrands), the p-integral benchmark, the Comparison Test, and fully worked examples at every stage.
What Are Improper Integrals? (Definition)
Now that we have a sense of why these integrals matter, let us pin down exactly what makes an integral “improper.”
An integral \(\displaystyle\int_a^b f(x)\,dx\) is called an improper integral if at least one of the following conditions holds:
- One or both limits of integration are infinite (i.e., \(a = -\infty\) or \(b = +\infty\), or both).
- The integrand \(f(x)\) is unbounded (tends to \(\pm\infty\)) at one or more points in the interval \([a, b]\).
In either case, the integral is defined as a limit of proper integrals — integrals over finite intervals of bounded functions. If that limit exists and is a finite real number, the improper integral is said to converge. If the limit fails to exist or is infinite, the improper integral is said to diverge.
Plain-English translation: you cannot directly “plug in” infinity as a number or evaluate a function at a point where it has no finite value. Instead, you approach the troublesome boundary using a limit — you integrate up to some finite value \(t\), find the answer in terms of \(t\), and then let \(t\) move towards the boundary.
⚠️ Common mistake: Students sometimes treat \(\infty\) as an ordinary number and write \(F(\infty) – F(a)\) using the antiderivative \(F\). This is not valid. Infinity is not in the domain of any real-valued function. You must always convert the infinite boundary into a limit first, evaluate the antiderivative at the finite upper limit \(t\), and only then take \(t \to \infty\).
Type I Improper Integrals: Infinite Limits of Integration
The first class of improper integrals — and the one most students encounter earliest — arises when at least one of the limits of integration is infinite.
Formal Definitions: Type I
Case 1 — Infinite upper limit. If \(f\) is continuous on \([a, +\infty)\), then:
\int_a^{+\infty} f(x)\,dx \;=\; \lim_{t \to +\infty} \int_a^t f(x)\,dx
\]
Case 2 — Infinite lower limit. If \(f\) is continuous on \((-\infty, b]\), then:
\int_{-\infty}^{b} f(x)\,dx \;=\; \lim_{t \to -\infty} \int_t^b f(x)\,dx
\]
Case 3 — Both limits infinite. If \(f\) is continuous on \((-\infty, +\infty)\), choose any convenient splitting point \(c\) (often \(c = 0\)), then:
\int_{-\infty}^{+\infty} f(x)\,dx \;=\; \int_{-\infty}^{c} f(x)\,dx \;+\; \int_{c}^{+\infty} f(x)\,dx
\]
The doubly-infinite integral converges only if both component integrals converge independently. The splitting point \(c\) is arbitrary — the answer is the same regardless of which value you choose.
⚠️ Common mistake: For a doubly-infinite integral \(\int_{-\infty}^{+\infty} f(x)\,dx\), it is not valid to write \(\lim_{t\to\infty}\int_{-t}^{t} f(x)\,dx\) and call that the answer. This symmetric limit (called the Cauchy principal value) can be finite even when the true improper integral diverges. The two halves must converge separately.
Worked Example: A Convergent Type I Integral
Evaluate \(\displaystyle\int_1^{+\infty} \frac{1}{x^2}\,dx\).
Step 1 — Replace the infinite limit with a variable.
\int_1^{+\infty} \frac{1}{x^2}\,dx \;=\; \lim_{t \to +\infty} \int_1^t x^{-2}\,dx
\]
Step 2 — Evaluate the antiderivative on the finite interval. The antiderivative of \(x^{-2}\) is \(-x^{-1} = -\dfrac{1}{x}\).
\lim_{t \to +\infty} \left[-\frac{1}{x}\right]_1^t \;=\; \lim_{t \to +\infty} \left(-\frac{1}{t} + 1\right)
\]
Step 3 — Take the limit. As \(t \to +\infty\), the term \(-\tfrac{1}{t}\) tends to 0:
\lim_{t \to +\infty} \left(1 – \frac{1}{t}\right) \;=\; 1
\]
The integral converges to \(1\). Geometrically, the region between the curve \(y = 1/x^2\) and the \(x\)-axis, stretching infinitely to the right starting from \(x = 1\), encloses a finite area of exactly 1 square unit — a beautifully counter-intuitive result.
Worked Example: A Divergent Type I Integral
Evaluate \(\displaystyle\int_1^{+\infty} \frac{1}{x}\,dx\). This is the harmonic integral, the continuous analogue of the famous harmonic series, and understanding its relationship to the definite integral is central to the convergence of p-series.
\lim_{t \to +\infty} \int_1^t \frac{1}{x}\,dx \;=\; \lim_{t \to +\infty} \bigl[\ln x\bigr]_1^t \;=\; \lim_{t \to +\infty} \ln t \;=\; +\infty
\]
The limit is infinite, so the integral diverges. Even though \(1/x\) shrinks towards zero as \(x\) grows, it shrinks too slowly to keep the cumulative area finite.
Type II Improper Integrals: Discontinuous Integrands
The second major type looks perfectly ordinary at first glance — the limits of integration are finite — but hides a singularity (a point where the function becomes infinite) somewhere in the interval.
Formal Definitions: Type II
Case 1 — Discontinuity at the upper limit. If \(f\) is continuous on \([a, b)\) but \(\lim_{x \to b^-} f(x) = \pm\infty\), then:
\int_a^b f(x)\,dx \;=\; \lim_{t \to b^-} \int_a^t f(x)\,dx
\]
Case 2 — Discontinuity at the lower limit. If \(f\) is continuous on \((a, b]\) but \(\lim_{x \to a^+} f(x) = \pm\infty\), then:
\int_a^b f(x)\,dx \;=\; \lim_{t \to a^+} \int_t^b f(x)\,dx
\]
Case 3 — Discontinuity in the interior. If \(f\) has an infinite discontinuity at an interior point \(c \in (a, b)\), split the integral:
\int_a^b f(x)\,dx \;=\; \int_a^c f(x)\,dx \;+\; \int_c^b f(x)\,dx
\]
Both pieces must converge for the whole integral to converge.
Worked Example: Type II Integral
Evaluate \(\displaystyle\int_0^1 \frac{1}{\sqrt{x}}\,dx\).
Here \(f(x) = x^{-1/2}\) is continuous on \((0, 1]\) but blows up as \(x \to 0^+\), so this is a Type II improper integral with a discontinuity at the lower limit.
\begin{align*}
\int_0^1 \frac{1}{\sqrt{x}}\,dx &= \lim_{t \to 0^+} \int_t^1 x^{-1/2}\,dx \\[6pt]
&= \lim_{t \to 0^+} \Bigl[2\sqrt{x}\Bigr]_t^1 \\[6pt]
&= \lim_{t \to 0^+} \bigl(2 – 2\sqrt{t}\bigr) \\[6pt]
&= 2 – 0 \;=\; 2
\end{align*}
\]
The integral converges to \(2\). Even though the function shoots off to infinity at \(x = 0\), the area it encloses remains finite.
⚠️ Common mistake: The integral \(\displaystyle\int_{-1}^{1} \frac{1}{x^2}\,dx\) looks like a standard definite integral, but \(f(x) = 1/x^2\) has an infinite discontinuity at \(x = 0\), right in the middle of \([-1,1]\). Applying the Fundamental Theorem of Calculus blindly gives the nonsensical answer \(-2\). You must split the integral at \(x = 0\) and treat each piece as a Type II improper integral. Both pieces diverge, so the whole integral diverges.
The p-Integral: A Key Benchmark for Convergence
Before diving into convergence tests, it pays to know one family of improper integrals — the p-integrals — almost by heart, because they serve as the reference standard against which all other integrals are compared.
The p-Integral Theorem
For a real constant \(p\), the improper integral
\int_1^{+\infty} \frac{1}{x^p}\,dx
\]
- converges to \(\dfrac{1}{p-1}\) if \(p > 1\);
- diverges if \(p \leq 1\).
Similarly, for the singularity version at \(x = 0\):
\int_0^{1} \frac{1}{x^p}\,dx \quad \text{converges if } p < 1; \quad \text{diverges if } p \geq 1. \]
Notice the directions are reversed for the two cases: the integral at infinity likes large decay (needs \(p > 1\)), while the integral near a singularity likes mild blowup (needs \(p < 1\)). The borderline case \(p = 1\) — which gives \(\int 1/x\,dx = \ln x\) — always diverges. This is directly connected to understanding infinite series and their convergence via the p-series test.
Quick Proof for \(p \neq 1\)
\begin{align*}
\int_1^{+\infty} \frac{1}{x^p}\,dx &= \lim_{t\to+\infty} \int_1^t x^{-p}\,dx \\[6pt]
&= \lim_{t\to+\infty} \left[\frac{x^{1-p}}{1-p}\right]_1^t \\[6pt]
&= \lim_{t\to+\infty} \frac{t^{1-p} – 1}{1-p}
\end{align*}
\]
If \(p > 1\), then \(1 – p < 0\), so \(t^{1-p} \to 0\) as \(t \to +\infty\), giving the finite value \(\dfrac{1}{p-1}\).
If \(p < 1\), then \(1 - p > 0\), so \(t^{1-p} \to +\infty\), and the integral diverges.
If \(p = 1\), use \(\ln x\) as the antiderivative and the limit is again \(+\infty\).
Conditions for Convergence and Divergence
Knowing the p-integral by heart is only the beginning — the deeper skill is deciding whether an integral you have never seen before converges or diverges, especially when direct evaluation is difficult or impossible.
The Comparison Test for Improper Integrals
Suppose \(f(x)\) and \(g(x)\) are continuous on \([a, +\infty)\) and satisfy \(0 \leq f(x) \leq g(x)\) for all \(x \geq a\). Then:
- If \(\displaystyle\int_a^{+\infty} g(x)\,dx\) converges, then \(\displaystyle\int_a^{+\infty} f(x)\,dx\) also converges.
- If \(\displaystyle\int_a^{+\infty} f(x)\,dx\) diverges, then \(\displaystyle\int_a^{+\infty} g(x)\,dx\) also diverges.
Intuitively: if a smaller function’s area is already infinite, the larger function’s area must also be infinite. And if a larger function’s area is finite, anything trapped below it must also be finite.
Comparison Test: Worked Example
Determine whether \(\displaystyle\int_1^{+\infty} \frac{\cos^2 x}{x^2}\,dx\) converges or diverges.
Direct evaluation is awkward because \(\cos^2 x\) oscillates. But we know that \(0 \leq \cos^2 x \leq 1\) for all \(x\), so:
0 \;\leq\; \frac{\cos^2 x}{x^2} \;\leq\; \frac{1}{x^2}
\]
We know from the p-integral theorem (with \(p = 2 > 1\)) that \(\int_1^{+\infty} \frac{1}{x^2}\,dx\) converges. Since our function is trapped between 0 and the convergent \(1/x^2\), the Comparison Test guarantees that \(\int_1^{+\infty} \frac{\cos^2 x}{x^2}\,dx\) also converges.
The Limit Comparison Test
The Limit Comparison Test is particularly useful when the direct comparison is hard to establish but the two functions behave similarly for large \(x\). If \(f(x) \geq 0\) and \(g(x) > 0\) for \(x \geq a\), and
\lim_{x \to +\infty} \frac{f(x)}{g(x)} = L, \quad 0 < L < +\infty, \]
then \(\int_a^{+\infty} f(x)\,dx\) and \(\int_a^{+\infty} g(x)\,dx\) both converge or both diverge.
Geometric Intuition: What Improper Integrals Look Like
Understanding improper integrals analytically is one thing, but holding a clear picture of them in your mind makes every calculation more natural.
For a Type I integral like \(\int_1^{+\infty} \frac{1}{x^2}\,dx\), imagine the region between the curve \(y = 1/x^2\) and the positive \(x\)-axis. The curve starts at \(y = 1\) when \(x = 1\) and descends steeply towards zero as \(x\) grows. The key geometric insight is that even though this region extends infinitely to the right, it also narrows fast enough that the total area remains bounded.
Contrast this with \(\int_1^{+\infty} \frac{1}{x}\,dx\): the curve \(y = 1/x\) also approaches zero, but it does so more slowly — slowly enough that the area beneath it accumulates without bound.
For a Type II integral like \(\int_0^1 \frac{1}{\sqrt{x}}\,dx\), the region is bounded on the right at \(x = 1\), but it shoots upward to infinity at \(x = 0\). The curve has a vertical asymptote — a line it approaches but never touches — at the left edge of the interval. Yet the area enclosed between this spike and the \(x\)-axis is only 2, a finite number, because the spike is so thin near the axis.
This geometric viewpoint also explains why understanding limits and continuity is the essential prerequisite: the entire method rests on approaching a boundary that cannot be reached directly.
Step-by-Step Proof: Convergence of the p-Integral for \(p > 1\)
With the geometric picture and the comparison tools in place, let us now work through a fully rigorous derivation of the most important result in this topic.
Claim: For \(p > 1\), \(\displaystyle\int_1^{+\infty} x^{-p}\,dx = \dfrac{1}{p-1}\).
Step 1 — Recognise the improper nature.
The upper limit is \(+\infty\), which is not a real number. We therefore write the integral as a limit. Let \(t > 1\) be a finite real number:
\int_1^{+\infty} x^{-p}\,dx \;=\; \lim_{t \to +\infty} \int_1^t x^{-p}\,dx
\]
Step 2 — Evaluate the proper integral using the Power Rule.
Since \(p \neq 1\), the antiderivative of \(x^{-p}\) is \(\dfrac{x^{-p+1}}{-p+1} = \dfrac{x^{1-p}}{1-p}\).
\int_1^t x^{-p}\,dx \;=\; \left[\frac{x^{1-p}}{1-p}\right]_1^t \;=\; \frac{t^{1-p}}{1-p} – \frac{1}{1-p} \;=\; \frac{t^{1-p} – 1}{1-p}
\]
Step 3 — Analyse the behaviour of \(t^{1-p}\) as \(t \to +\infty\).
Since \(p > 1\), we have \(1 – p < 0\). An expression \(t^{\alpha}\) with negative exponent \(\alpha\) satisfies \(t^{\alpha} \to 0\) as \(t \to +\infty\).
Step 4 — Take the limit.
\lim_{t \to +\infty} \frac{t^{1-p} – 1}{1-p} \;=\; \frac{0 – 1}{1-p} \;=\; \frac{-1}{1-p} \;=\; \frac{1}{p-1}
\]
Step 5 — Conclude.
The limit exists and is the finite number \(\dfrac{1}{p-1}\). Therefore the improper integral converges, and its value is \(\dfrac{1}{p-1}\). \(\square\)
As a quick check: with \(p = 2\) the formula gives \(\tfrac{1}{2-1} = 1\), which matches our earlier direct calculation. With \(p = 3\) it gives \(\tfrac{1}{2}\), and so on.
Conclusion
Improper integrals are a natural and powerful extension of standard integration, enabling calculus to handle the infinite — whether that means integrating over an unbounded interval or integrating a function with a vertical asymptote. The central technique in every case is the same: replace the problematic boundary with a finite parameter \(t\), integrate properly, then take the limit. When that limit is a finite number, the improper integral converges; when it is infinite or does not exist, the integral diverges.
The most important benchmark to know is the p-integral: \(\int_1^{+\infty} x^{-p}\,dx\) converges precisely when \(p > 1\). Every comparison test argument ultimately comes back to this touchstone result. When direct evaluation is not feasible, the Comparison Test and Limit Comparison Test give you the tools to decide convergence without ever computing an explicit antiderivative.