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.

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.
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
For an infinite lower tail, use
If both bounds are infinite, split at a finite point 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 , approach it from inside the interval:
If the upper endpoint is unbounded, approach from the left:
If an interior point is unbounded, with , test both one-sided pieces:
The integral converges to 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.
| Integral | Finite when | Value |
|---|---|---|
At infinity, gives divergence. Near zero, gives divergence. For , the near-zero integral is improper but convergent. If , 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 , the exponential integrand is ; for , 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.
Choose a pattern and check its convergence.
5. Eight worked examples with complete limit setups
Example 1 · Evaluate a convergent infinite upper tail
Evaluate .
1 · Replace the infinite bound by a finite cutoff
The integrand is continuous on every finite interval .
2 · Evaluate the finite integral first
A primitive is .
3 · Take the limit and state convergence
The required limit is finite, so the improper integral converges to .
Example 2 · A function can approach zero while its integral diverges
Determine whether converges.
1 · Use a finite cutoff
2 · Evaluate the logarithmic limit
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
A divergent tail that also tends to zero
Example 3 · Split an integral with two infinite tails
Evaluate .
1 · Test the left tail
Split at , and use the primitive .
2 · Test the right tail independently
3 · Add only after both limits are finite
Both tails converge, so the original improper integral converges to .
Example 4 · Approach an unbounded lower endpoint from the right
Evaluate .
1 · Keep the finite cutoff inside the domain
Use , because the original integrand is defined only for .
2 · Evaluate and take the one-sided limit
The integral converges, despite the unbounded height near zero.
Example 5 · Do not cancel divergence at an interior pole
Determine whether converges.
1 · Test the left side of the pole
The pole at lies inside the interval. For , the finite left integral has value .
2 · Test the right side separately
For , the integral from to is .
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
Example 6 · Approach an unbounded upper endpoint from the left
Evaluate .
1 · Place the cutoff before the endpoint
For , use the primitive .
2 · Take the required one-sided limit
The integral converges to .
Example 7 · Handle an infinite lower bound
Evaluate .
1 · Start at a finite point on the left
2 · Evaluate before passing to the limit
Because as , the integral converges to .
Example 8 · Split a problem with two different improper features
Evaluate .
1 · Identify both issues and a primitive
The integrand is unbounded at and has an infinite upper tail. Split at . On , a primitive is
It follows by substituting , or by differentiating .
2 · Take the lower-endpoint limit
3 · Take the tail limit independently
Both pieces are finite, so this improper integral converges to .
6. Practice with step-by-step solutions
Problem 1
Evaluate .
Show solution 1
Solution
A primitive is . Evaluate from to a finite cutoff .
The integral converges.
Problem 2
Determine whether converges.
Show solution 2
Solution
The exponent is below the tail threshold. Direct evaluation gives at cutoff , which grows without bound as . The improper integral diverges.
Problem 3
Evaluate .
Show solution 3
Solution
Use the lower-endpoint limit from the right. The exponent is below .
The improper integral converges.
Problem 4
Determine whether converges.
Show solution 4
Solution
With , the finite integral is .
The improper integral diverges.
Problem 5
Evaluate .
Show solution 5
Solution
The finite-cutoff integral is .
The integral converges.
Problem 6
Evaluate .
Show solution 6
Solution
Split at . On the right, the one-sided integral equals , as in Example 4. On the left, substitute and evaluate its separate limit; that piece also equals .
An interior unbounded point does not force divergence; both required limits are finite here.
Problem 7
Determine whether converges.
Show solution 7
Solution
Split at . A primitive is . The left finite piece from to is , which tends to as . The right tail tends to . Therefore the improper integral diverges. A symmetric cancellation is not a convergence test.
Problem 8
Determine whether converges.
Show solution 8
Solution
The cutoff integral is . At it is , while at it is . As positive integers 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
- Inspect the interval and integrand before evaluating. Identify every infinite end and unbounded point.
- Write finite-cutoff integrals and the correct limit directions. Split both infinite tails and every interior singularity.
- Evaluate each ordinary definite integral before taking its limit. Infinity is not a number to substitute into a primitive.
- Decide whether every required limit is finite. Infinite limits and oscillating limits both give divergence.
- 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 -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 converges for . The near-zero integral is finite for ; it is genuinely unbounded at zero only when .
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.