AP® CALCULUS BC · UNIT 6

Evaluating Improper Integrals

Use limits to handle infinite intervals and unbounded integrands. Compare convergent and divergent examples, test common patterns, and justify every conclusion.

Topic 6.13 study notes8 worked examples · 8 practice questionsUpdated 11 November 2026
Calculus notebook with limit and integral symbols, a compass and pencil on an olive and cream study desk

1. What makes an integral improper?

An improper integral has an infinite integration bound, an integrand that becomes unbounded on the interval, or both. Replace each problematic boundary with a limit of ordinary definite integrals.

AP Calculus BC only · Topic 6.13. Evaluate the required limits and decide whether the improper integral converges to a finite real value or diverges.

An unbounded region can have finite area. A finite jump or removable discontinuity alone does not make a bounded function's integral improper. Identify the actual issue rather than labeling every discontinuity a vertical asymptote.

2. Write the required limits and split points

Infinite intervals: test each tail independently

For a continuous integrand on each finite interval, define an infinite upper tail by

∫a∞f(x)dx=limb→∞∫abf(x)dx.

For an infinite lower tail, use

∫−∞bf(x)dx=lima→−∞∫abf(x)dx.

If both bounds are infinite, split at a finite point c where the integrand is continuous. Test the left and right tails separately. Add them only when both are finite; one symmetric cutoff can conceal divergence.

Unbounded integrands: use the correct approach direction

If the integrand is unbounded at the lower endpoint a, approach it from inside the interval:

∫abf(x)dx=limt→a+∫tbf(x)dx.

If the upper endpoint b is unbounded, approach from the left:

∫abf(x)dx=limt→b−∫atf(x)dx.

If an interior point c is unbounded, with a<c<b, test both one-sided pieces:

L=limt→c−∫atf(x)dx
R=lims→c+∫sbf(x)dx

The integral converges to L+R only if both limits are finite. Split at every unbounded point or infinite end when a problem has more than one issue.

3. Power and exponential convergence rules

The two power patterns have opposite thresholds. At infinity the tail must decay fast enough; near zero, an unbounded power must not grow too strongly.

Power and exponential patterns
IntegralFinite whenValue
∫1∞x−pdxp>11p−1
∫01x−pdxp<111−p
∫0∞e−kxdxk>01k

At infinity, p≤1 gives divergence. Near zero, p≥1 gives divergence. For 0<p<1, the near-zero integral is improper but convergent. If p≤0, the power has a bounded continuous extension at zero, so the near-zero case is a proper finite integral after that extension. The same value formula applies.

For k=0, the exponential integrand is 1; for k<0, it grows. Both exponential tails diverge. These rules follow from antiderivatives and limits, rather than from the function's approaching zero alone.

4. Interactive convergence checker

Explore the original four patterns. This checker handles these standard families; use the worked examples for full limit notation and for integrands with several improper points.

Enter a finite real parameter. The interior pattern needs a separate limit on each side and does not use this field.

∫1∞x−pdx

Choose a pattern and check its convergence.

5. Eight worked examples with complete limit setups

Example 1 · Evaluate a convergent infinite upper tail

Evaluate I=∫1∞1x2dx.

1 · Replace the infinite bound by a finite cutoff

I=limb→∞∫1bx−2dx

The integrand is continuous on every finite interval [1,b].

2 · Evaluate the finite integral first

∫1bx−2dx=1−1b

A primitive is −1x.

3 · Take the limit and state convergence

I=limb→∞(1−1b)=1

The required limit is finite, so the improper integral converges to 1.

Example 2 · A function can approach zero while its integral diverges

Determine whether I=∫1∞1xdx converges.

1 · Use a finite cutoff

I=limb→∞∫1b1xdx

2 · Evaluate the logarithmic limit

∫1b1xdx=lnb
limb→∞lnb=+∞

The limit is not finite, so the improper integral diverges. Approaching zero is not enough to guarantee a finite tail area.

A convergent infinite tail

image/svg+xml Matplotlib v3.11.2, https://matplotlib.org/ 1 2 4 6 8 x 0.00 0.25 0.50 0.75 1.00 f x ( )
Only [1,8] is shaded for f(x)=1x2. Its area is 78, and the remaining tail has area 18. The total improper integral is 1.

A divergent tail that also tends to zero

image/svg+xml Matplotlib v3.11.2, https://matplotlib.org/ 1 2 4 6 8 x 0.00 0.25 0.50 0.75 1.00 g x ( )
Only [1,8] is shaded for g(x)=1x. Its area is ln8. The accumulated area keeps growing as the cutoff increases, even though g(x)→0.

Example 3 · Split an integral with two infinite tails

Evaluate I=∫−∞∞11+x2dx.

1 · Test the left tail

Split at 0, and use the primitive arctanx.

L=lima→−∞[arctanx]a0
L=π2

2 · Test the right tail independently

R=limb→∞[arctanx]0b
R=π2

3 · Add only after both limits are finite

I=L+R=π

Both tails converge, so the original improper integral converges to π.

Example 4 · Approach an unbounded lower endpoint from the right

Evaluate I=∫011xdx.

1 · Keep the finite cutoff inside the domain

I=limt→0+∫t1x−12dx

Use t>0, because the original integrand is defined only for x>0.

2 · Evaluate and take the one-sided limit

∫t1x−12dx=2−2t
I=limt→0+(2−2t)=2

The integral converges, despite the unbounded height near zero.

Example 5 · Do not cancel divergence at an interior pole

Determine whether I=∫021x−1dx converges.

1 · Test the left side of the pole

The pole at 1 lies inside the interval. For t<1, the finite left integral has value ln(1−t).

L=limt→1−ln(1−t)=−∞

2 · Test the right side separately

For s>1, the integral from s to 2 is −ln(s−1).

R=lims→1+[−ln(s−1)]=+∞

3 · State divergence before any addition

Neither one-sided integral is finite, so the original improper integral diverges. Equal cutoffs on the two sides give a symmetric cancellation, sometimes called a principal value, but that is not the ordinary improper integral.

Finite cutoffs can hide a divergent interior pole

image/svg+xml Matplotlib v3.11.2, https://matplotlib.org/ 0.0 0.5 1.0 1.5 2.0 x −8 −4 0 4 8 f x ( )
The shaded pieces stop at 34 and restart at 54. Their signed values are −ln4 and ln4. Moving these cutoffs independently toward 1 makes the two limits diverge.

Example 6 · Approach an unbounded upper endpoint from the left

Evaluate I=∫0111−xdx.

1 · Place the cutoff before the endpoint

I=limt→1−∫0t11−xdx

For t<1, use the primitive −21−x.

2 · Take the required one-sided limit

∫0t11−xdx=2−21−t
I=limt→1−(2−21−t)=2

The integral converges to 2.

Example 7 · Handle an infinite lower bound

Evaluate I=∫−∞0exdx.

1 · Start at a finite point on the left

I=lima→−∞∫a0exdx

2 · Evaluate before passing to the limit

I=lima→−∞(1−ea)=1

Because ea→0 as a→−∞, the integral converges to 1.

Example 8 · Split a problem with two different improper features

Evaluate I=∫0∞1x(1+x)dx.

1 · Identify both issues and a primitive

The integrand is unbounded at 0 and has an infinite upper tail. Split at 1. On x>0, a primitive is

F(x)=2arctanx.

It follows by substituting u=x, or by differentiating F.

2 · Take the lower-endpoint limit

A(t)=F(1)−F(t)
L=limt→0+A(t)=π2

3 · Take the tail limit independently

B(b)=F(b)−F(1)
R=limb→∞B(b)=π2
I=L+R=π

Both pieces are finite, so this improper integral converges to π.

6. Practice with step-by-step solutions

Problem 1

Evaluate I=∫1∞1x3dx.

Show solution 1

Solution

A primitive is −12x2. Evaluate from 1 to a finite cutoff b.

I=limb→∞(12−12b2)=12

The integral converges.

Problem 2

Determine whether I=∫1∞x−45dx converges.

Show solution 2

Solution

The exponent p=45=0.8 is below the tail threshold. Direct evaluation gives 5(b15−1) at cutoff b, which grows without bound as b→∞. The improper integral diverges.

Problem 3

Evaluate I=∫01x−13dx.

Show solution 3

Solution

Use the lower-endpoint limit from the right. The exponent is below 1.

I=limt→0+32(1−t23)=32

The improper integral converges.

Problem 4

Determine whether I=∫011xdx converges.

Show solution 4

Solution

With t>0, the finite integral is −lnt.

limt→0+(−lnt)=+∞

The improper integral diverges.

Problem 5

Evaluate I=∫0∞e−2xdx.

Show solution 5

Solution

The finite-cutoff integral is 12(1−e−2b).

I=limb→∞12(1−e−2b)=12

The integral converges.

Problem 6

Evaluate I=∫−111|x|dx.

Show solution 6

Solution

Split at 0. On the right, the one-sided integral equals 2, as in Example 4. On the left, substitute u=−x and evaluate its separate limit; that piece also equals 2.

I=2+2=4

An interior unbounded point does not force divergence; both required limits are finite here.

Problem 7

Determine whether I=∫−∞∞x1+x2dx converges.

Show solution 7

Solution

Split at 0. A primitive is 12ln(1+x2). The left finite piece from a to 0 is −12ln(1+a2), which tends to −∞ as a→−∞. The right tail tends to +∞. Therefore the improper integral diverges. A symmetric cancellation is not a convergence test.

Problem 8

Determine whether I=∫0∞sinxdx converges.

Show solution 8

Solution

The cutoff integral is 1−cosb. At b=2nπ it is 0, while at b=(2n+1)π it is 2. As positive integers n grow, these values do not approach the same limit. The improper integral diverges because its limit does not exist.

7. A reliable process and common mistakes

  1. Inspect the interval and integrand before evaluating. Identify every infinite end and unbounded point.
  2. Write finite-cutoff integrals and the correct limit directions. Split both infinite tails and every interior singularity.
  3. Evaluate each ordinary definite integral before taking its limit. Infinity is not a number to substitute into a primitive.
  4. Decide whether every required limit is finite. Infinite limits and oscillating limits both give divergence.
  5. End with a conclusion: converges to the finite value, or diverges with the failed limit identified.

Keep absolute values in logarithmic primitives until the interval establishes the sign. Do not cancel opposing infinite limits or infer convergence merely because the integrand approaches zero. The finite p-integral rules depend on which endpoint is being tested.

8. Questions about convergence and divergence

What makes an integral improper?

An infinite integration bound or an integrand that becomes unbounded on the interval requires an improper limit. A bounded jump or removable discontinuity alone does not create this issue.

How do you evaluate an improper integral?

Write the required limit or separate limits, evaluate the finite definite integral expressions, then test the limits. Give the finite value only when every piece converges.

What does convergence mean?

Every required limiting integral has a finite real value. When there are several pieces, their finite values can then be added.

What counts as divergence?

A required limit is infinite or does not exist. An oscillating cutoff integral can diverge even without tending to infinity.

When do the power patterns converge?

The tail ∫1∞x−pdx converges for p>1. The near-zero integral ∫01x−pdx is finite for p<1; it is genuinely unbounded at zero only when p>0.

Why split when both ends are infinite?

Each tail must converge separately. A limit with one symmetric cutoff can cancel opposing divergence and give a principal value instead of the ordinary improper integral.

Why split an interior unbounded point?

It creates separate one-sided integrals. The original integral converges only if both are finite, regardless of whether symmetric cutoffs happen to cancel.

9. Continue studying integration

Review Topic 6.7's finite endpoint evaluation inside each limit. Topic 6.12's partial fractions can help find a primitive, but you must still test the improper limits.