An improper integral is a definite integral where the interval is infinite (Type I) or the integrand blows up somewhere in the interval (Type II); you evaluate it as a limit of proper integrals and ask whether that limit is finite (converges) or not (diverges).
Type I: ∫a∞f(x)dx=limb→∞∫abf(x)dx Type II: ∫abf(x)dx=limc→a+∫cbf(x)dx (if f is unbounded at a)
Quick Answer
An improper integral is a definite integral where the interval is infinite (Type I) or the integrand blows up somewhere in the interval (Type II); you evaluate it as a limit of proper integrals and ask whether that limit is finite (converges) or not (diverges). Use it for totals over unlimited time, such as all the water a slow leak will ever lose, or for areas beside a vertical asymptote, such as ∫01x1dx. Replace the problem point by a variable, integrate, then take the limit.
Why This Matters
Probability, physics and economics keep needing totals over endless ranges: the chance a waiting time exceeds 15 minutes, the energy to launch a rocket free of Earth's gravity, the value of income that never stops. Improper integrals make such totals precise and show when they are finite. Treating ∞ as a number, or integrating straight across an asymptote, produces results like ∫−11x21dx=−2, a negative area for a positive function.
Intuitive Explanation
The region under y=e−x for x≥0 is infinitely long, yet it has a finite area. Cut it off at x=b: ∫0be−xdx=1−e−b. That is about 0.865 at b=2, 0.993 at b=5 and 0.99995 at b=10. The totals approach 1, so we define ∫0∞e−xdx=limb→∞(1−e−b)=1. The curve drops so fast that each extra stretch adds less and less.
Power curves are the delicate case. ∫1bx21dx=1−b1 approaches 1, but ∫1bx1dx=lnb, about 2.3 at b=10 and 4.6 at b=100, grows without limit. The two graphs look alike; what differs is how fast they shrink. In general ∫1∞xp1dx converges exactly when p>1, the same dividing line as for the series ∑np1.
The second kind has a finite interval but an integrand that blows up. ∫01x1dx=limc→0+∫c1x−1/2dx=limc→0+(2−2c)=2: a finite area beside an infinitely tall wall. The danger is an asymptote hidden inside the interval. ∫−11x21dx looks like an ordinary definite integral, but x21 blows up at x=0; split the integral there and each half diverges.
Think of it like this.Pour paint into an endless trough whose depth keeps shrinking. Whether one can of paint can fill it depends on how fast the trough gets shallow: depth x21 beyond x=1 takes exactly one unit of paint, while depth x1 swallows any amount. A real trough would end somewhere; this one never does, which is why you take a limit.
Core idea.Integrate over an infinite interval or through a blow-up by taking a limit of ordinary integrals.
When to Use
Use it when a definite integral has ∞ or −∞ as a limit, or when the integrand has a vertical asymptote at an endpoint or inside the interval; check for zeros of a denominator before integrating. Replace the problem point by a variable and take a limit. For a proper integral, with finite limits and a bounded integrand, the Fundamental Theorem applies directly. For an infinite sum of terms rather than an integral, use series convergence tests.
Tip.Ask: Does this integral run to infinity or pass through a point where the integrand blows up?
How to Recognize It
Inspect both limits of integration and every point in between, including zeros of a denominator, before you write down an antiderivative.
1.
Does a denominator vanish somewhere strictly inside the interval?
Yes means split the integral at that point into two improper pieces, and each must converge. Ignoring the break gives nonsense: ∫−11x21dx is not −2; it diverges.
2.
Is the question about a sum of terms ∑an rather than an area?
A sum of separate terms is series convergence. The two are linked by the integral test: ∑np1 and ∫1∞xp1dx converge for exactly the same p, namely p>1.
3.
Does a positive integrand shrink at least as fast as xp1 for some p>1, like x21 or e−x?
Yes means the infinite tail converges, and a comparison often settles it without an antiderivative. An integrand that shrinks like x1 or more slowly, such as x1, gives an infinite area.
4.
Does the problem describe a total over unlimited time, like all the medicine a body will ever clear?
Yes is an infinite upper limit in disguise: ∫0∞5e−0.2tdt=25. A total over a fixed span, such as the first 10 hours, is an ordinary proper definite integral.
Improper Integrals compared with Proper definite integral, Series convergence, and Indeterminate-form limit
All four rows involve a limit and the question of whether a total is finite, so they appear side by side. Ask what is being totalled, an area or a list of terms, and where the limit is taken: at a bound of the integral or inside the function.
Improper Integrals compared with Proper definite integral, Series convergence, and Indeterminate-form limit
Type
Meaning
Key test
Formula
Example
Improper Integrals
Use this when a definite integral has an infinite limit of integration or an infinite discontinuity in the integrand.
Does this integral run to infinity or pass through a point where the integrand blows up?
Type I: ∫a∞f(x)dx=limb→∞∫abf(x)dx Type II: ∫abf(x)dx=limc→a+∫cbf(x)dx (if f is unbounded at a)
Evaluate ∫1∞x21dx.
Proper definite integral
Has finite bounds and a bounded integrand; no limit needed.
Use when the interval is finite and the function stays bounded.
∫abfdx
∫02xdx=2
Series convergence
Decides if an infinite SUM (discrete) is finite; improper integrals do the same for an integral (continuous).
Use for ∑an; they are linked by the integral test.
∑an
∑1/n2 converges
Indeterminate-form limit
Resolves 00 etc. inside a function; improper integrals take a limit of the WHOLE integral.
Use L'Hopital for the integrand's form, then the bound-limit for the integral.
L'Hopital
limx→0xsinx
Improper Integrals
Meaning
Use this when a definite integral has an infinite limit of integration or an infinite discontinuity in the integrand.
Key test
Does this integral run to infinity or pass through a point where the integrand blows up?
Formula
Type I: ∫a∞f(x)dx=limb→∞∫abf(x)dx Type II: ∫abf(x)dx=limc→a+∫cbf(x)dx (if f is unbounded at a)
Example
Evaluate ∫1∞x21dx.
Proper definite integral
Meaning
Has finite bounds and a bounded integrand; no limit needed.
Key test
Use when the interval is finite and the function stays bounded.
Formula
∫abfdx
Example
∫02xdx=2
Series convergence
Meaning
Decides if an infinite SUM (discrete) is finite; improper integrals do the same for an integral (continuous).
Key test
Use for ∑an; they are linked by the integral test.
Formula
∑an
Example
∑1/n2 converges
Indeterminate-form limit
Meaning
Resolves 00 etc. inside a function; improper integrals take a limit of the WHOLE integral.
Key test
Use L'Hopital for the integrand's form, then the bound-limit for the integral.
Formula
L'Hopital
Example
limx→0xsinx
Formula & Notation
Type I: ∫a∞f(x)dx=limb→∞∫abf(x)dx Type II: ∫abf(x)dx=limc→a+∫cbf(x)dx (if f is unbounded at a)
Type I: ∫a∞f(x)dx=limb→∞∫abf(x)dx. Type II: if limx→a+f(x)=±∞, then ∫abf(x)dx=limc→a+∫cbf(x)dx. The integral converges if the limit is finite; it diverges otherwise.
How to read it:Type I: infinite interval (∫a∞, ∫−∞b, ∫−∞∞). Type II: infinite integrand (discontinuity inside [a,b]).
Worked Examples
Example 1 — Type I improper integral
Easy
Evaluate ∫1∞x21dx.
Show the steps and the answer
Replace the infinite bound: ∫1∞x21dx=limb→∞∫1bx−2dx.
∞ is not a number you can substitute, so integrate up to a finite b and let b grow.
Integrate: ∫1bx−2dx=[−x1]1b=−b1+1.
The antiderivative of x−2 is −x−1.
Take the limit: limb→∞(1−b1)=1.
b1→0, so the areas 0.9, 0.99, 0.999 for b=10,100,1000 settle on 1.
Answer:1 (converges)
Takeaway:Swap ∞ for b, integrate, then let b→∞; a finite limit means the integral converges.
Example 2 — Same family, diverges
Standard
Evaluate ∫1∞x1dx.
Show the steps and the answer
Set up the limit: limb→∞∫1bx1dx=limb→∞[lnx]1b=limb→∞lnb.
The method is the same as for x21, but the antiderivative is now a logarithm.
Watch lnb: about 2.3 at b=10, 6.9 at b=1000, 13.8 at b=106.
It grows slowly but without bound, so the limit is infinite.
Conclude: the integral diverges and has no finite value.
x1 and x21 look alike, but x1 shrinks too slowly; ∫1∞xp1dx converges only when p>1.
Answer:Diverges — no finite value
Takeaway:Two similar-looking curves can give opposite answers; only the limit decides convergence.
Example 3 — Find the mistake: a hidden asymptote
Application
Evaluate ∫−12x21dx. Ivy writes: "[−x1]−12=−21−1=−23." Find the mistake and correct it.
Show the steps and the answer
Notice the warning sign: a positive function cannot have a negative integral.
x21>0 everywhere it is defined, so −23 must be wrong.
Find the asymptote: x21 blows up at x=0, which lies inside [−1,2].
The Fundamental Theorem needs a continuous integrand, so Ivy's one-step evaluation is not allowed.
Split at 0 and test one piece: ∫02x21dx=limc→0+[−x1]c2=limc→0+(c1−21)=∞.
Each piece is its own improper integral, and this one has an infinite limit.
Conclude: the whole integral diverges.
If any piece diverges the integral has no finite value; the other piece, ∫−10x21dx, diverges too.
Answer:The integral diverges.
Takeaway:Scan the interval for points where the integrand blows up, split there, and treat each piece as an improper integral.
Where You'll See It
Escaping Earth's gravity
Lifting 1 kg from Earth's surface to infinitely far away takes work ∫R∞r2GMdr=RGM≈62.6 million joules: an infinite distance, but finite energy. Setting 21v2 equal to that gives the escape velocity, about 11.2 km/s.
Chance of a long wait for a bus
If buses come at random, on average every 10 minutes, the waiting time has density 101e−t/10. The chance of waiting more than 15 minutes is ∫15∞101e−t/10dt=e−1.5≈0.22.
All the water a leak will lose
A tank leaks at 20e−0.1t litres per minute, slowing as the pressure drops. The total it will ever lose is ∫0∞20e−0.1tdt=0.120=200 litres, even though the leak never quite stops.
Common Mistakes
The slip:Treating ∞ as a number to plug in
Instead:replace the infinite bound with a variable and take its limit explicitly.
The slip:Missing an interior asymptote
Instead:if the integrand blows up INSIDE [a,b], split the integral at that point (Type II), don't integrate across it.
The slip:Reporting an answer for a divergent integral
Instead:if the limit is infinite or fails to exist, the integral diverges (no finite value).
Mini Practice
Try these on your own. Tap Reveal when you want to check.
1.
Evaluate ∫1∞x31dx.
Hint:Replace ∞ with b, integrate, then let b→∞.
2.
Evaluate ∫0∞e−2xdx.
Hint:An antiderivative of e−2x is −21e−2x.
3.
Evaluate ∫01x1dx.
Hint:The integrand blows up at x=0. Start the integral at a and let a→0+.
4.
Evaluate ∫0∞1+x21dx.
Hint:An antiderivative is arctanx. What does it approach as x→∞?
5.
Is ∫02x−31dx an improper integral, or an ordinary definite integral?
Hint:Find where the integrand blows up and check whether that point lies in [0,2].
6.
A student computes "∫−11x21dx=[−x1]−11=−1−1=−2." Correct the work.
Hint:Is x21 defined everywhere on [−1,1]? Also, can a positive function have a negative integral?
Want the full set?
50 practice questions for this concept — free to try, every one with a complete worked
solution showing the why, not just the answer.
How can a shape have finite volume but infinite surface area?
Rotate y=x1 for x≥1 about the x-axis to get Gabriel's horn. Its volume is π∫1∞x21dx=π, but its surface area is at least 2π∫1∞x1dx, which diverges. So the horn holds π cubic units of paint, yet no amount of paint could coat its surface. Real paint has thickness, which is where the paradox breaks down.
Why isn't ∫−∞∞xdx equal to 0 by symmetry?
A two-sided improper integral must be split, say at 0, and each half must converge on its own. Here ∫0∞xdx diverges, so the whole integral diverges. Using ∫−bb with the same b on both sides hides this by letting two infinite areas cancel; with ∫−b2b you would get 23b2→∞, so the answer would depend on how you approach infinity.
Why can a rocket escape Earth's gravity with a finite amount of energy?
The work to lift a mass m from Earth's radius R to infinity is ∫R∞r2GMmdr=RGMm, a finite number, because gravity weakens like r21. Setting this equal to 21mv2 gives the escape speed R2GM, about 11.2 km/s for Earth. If gravity weakened only like r1, the integral would diverge and nothing could escape.
Can an improper integral have an exact value when the antiderivative has no formula?
Yes. The function e−x2 has no antiderivative built from familiar functions, yet ∫−∞∞e−x2dx=π, found with a clever trick that squares the integral and switches to polar coordinates. This result is the reason the normal distribution, the bell curve of statistics, has 2π in its formula.
You need Definite Integral, Limit and Infinity: an improper integral is the limit of ordinary definite integrals as a bound moves out to infinity or in toward an asymptote. It adds the idea that an unbounded region can have finite area. Next, Convergence and Divergence asks the same question of infinite series, and the integral test links the two: ∑np1 and ∫1∞xp1dx converge together.