Improper integrals with badly behaved limits are integrals where one or both of the limits is infinite.
These integrals look like

If only one limit of integration is infinite then the other limit of integration, c, would be a finite constant.
Remember that writing ± ∞ for the limits of integration is a shortcut. These improper integral creatures are really limits.
The integral

really means

Similarly,

To evaluate this type of improper integral, we have to work out the integral inside the limit and then evaluate the limit.
One Badly-Behaved Limit at a Time
When only one of the limits of integration is infinite we can work things out directly. When both limits of integration are infinite we have to split the integral into two pieces first.
Convergence and Divergence
There's some weird stuff going on here. If we take a value of p > 1, the integral

converges. But if we take a value of p < 1, that same integral diverges. And the graphs don't even look that different!
Take, for example, p = 2 and
:

One way to explain this is that the graph of
is able to bound a finite area above the x-axis on [1,∞) because it gets close to zero more quickly than the graph of
.
There's another thing going on here we need to point out. Let

Then the function f ( x ) converges to 0 as x approaches ∞. In symbols,

However, the integral

diverges. It's possible for the function to converge to 0 at ∞, but for an improper integral of the function, with an upper limit of ∞, to diverge.
However, if the function f ( x ) diverges at ∞, then there's no hope for the integral

to converge. No matter what c is, if f zooms off to ± ∞ or can't make up its mind what to equal as x approaches ∞, then

is going to diverge.
In fact, if

equals anything other than 0, even if the limit converges, then

will diverge. If
then the function and the x-axis will enclose an infinite area on the interval [c,∞).
With the other type of improper integral, it's possible for a function to diverge while the improper integral converges.
Be Careful: When you talk about the convergence or divergence of something, make sure you say what is doing the converging or diverging. Saying that a function converges and saying that an integral converges are not the same thing.
These ideas of convergence and divergence will come up again when we're talking about sequences and series.
Two Badly-Behaved Limits
So far we've looked at integrals with only one infinite limit. If both limits are infinite, we need to break up the integral somewhere in the middle.

In symbols,

where c is any real number you like (we usually pick something easy like 0 or 1, depending on the function). Since each of the integrals on the right-hand side is really a limit in disguise, this equation says

A sum of limits only exists if both limits involved exist. This is both good news and bad news.
The good news is that if you go to work out

and you find that either

diverges or

diverges, then you're done. The integral

diverges also.
The bad news is that if

does exist, you have to work out both of the limits

and

and then add them up. This comes with its own good news, though. You'll already have found the antiderivative for f ( x ) after the first limit, so the second limit won't be as much work.
Practice:
Determine whether 
converges or diverges, and find its value if it converges. | |
We've sort of done this problem already. We're pretty sure that this integral converges, and that it converges to 1. Let's argue the case a little more carefully. By the definition of this type of improper integral, we know that 
We know how to work out this integral, so we do - remembering to keep the limit notation along for the ride. 
We've worked out the integral inside the limit, so it's time to figure out the limit. As b approaches ∞, the term approaches 0. So 
The integral 
converges to 1. Remember to include the limit. If you leave a limit out, you get something like 
which doesn't make any sense. The left-hand side is a limit, which is either a number or doesn't exist at all. The right-hand side is just an algebraic expression with the letter b in it. | |
Determine if 
converges or diverges. Find its value if it converges. | |
We do pretty much the same stuff we did for the last problem. 
As b approaches -∞ the quantity approaches 0, so we have 
The function is even, so this answer makes sense. We're saying the area between and the x-axis on (-∞,-1) is the same as the area between and the x-axis on (1,∞). | |
Determine if 
converges or diverges. Indicate on a graph the region whose area is given by this integral. If the integral converges, find its value. | |
We write the improper integral as the limit it really is, work out the definite integral, then work out the limit. 
As b approaches ∞, the quantity ln b also approaches ∞. That means this limit doesn't exist, so we say the integral 
diverges. That means this region has infinite area. | |
Determine if 
converges or diverges. Find its value if it converges. | |
We need to split up the integral so we can deal with one badly-behaved limit of integration at a time. Let's split the integral at 0, since that's a nice easy number. 
Now we need to work out each of the simpler improper integrals (unless the first one diverges, in which case we'll be done!). 
Since ec2 approaches ∞ as c approaches -∞, the limit converges and 
The other improper integral is similar. 
Putting our results together, 
The original integral converges to 0. This makes sense, because in the graph it really looks like the total weighted area between the function and the x-axis should be 0. Yes, that was a lot of work to get 0 as an answer. Sometimes that's how it goes. And yes, it does make a lot of sense when you look at the graph that you would get zero for that integral. Alas, not all odd functions are that nice. The integral 
for example, doesn't exist. | |
Determine whether the integral converges or diverges. Indicate on a graph the region whose weighted area is given by the integral. If the integral converges, find its value.

Answer

As b approaches -∞ the quantity eb approaches 0, since a very negative power of e is the same thing as 1 divided by a very large power of e:

This means

This integral converges to 1.
Determine whether the integral converges or diverges. Indicate on a graph the region whose weighted area is given by the integral. If the integral converges, find its value.

Answer

This integral converges to
.
Determine whether the integral converges or diverges. Indicate on a graph the region whose weighted area is given by the integral. If the integral converges, find its value.

Answer

Since b1/2 approaches ∞ as b approaches ∞, this integral diverges.
Determine whether the integral converges or diverges. Indicate on a graph the region whose weighted area is given by the integral. If the integral converges, find its value.
for p > 1
Answer
Since we're told p ≠ 1, we know that the antiderivative of

is
(if p = 1 were an option, the antiderivative could be ln x instead). So when we work out the integral, we get

Since p > 1, the exponent p – 1 is greater than 0. As b approaches ∞ the quantity bp – 1 will grow larger and larger, so

will approach zero. This means the integral will converge to

Since p > 1 this quantity is positive, which is reassuring since the integral refers to an area above the x-axis.
Determine whether the integral converges or diverges. Indicate on a graph the region whose weighted area is given by the integral. If the integral converges, find its value.
for p < 1
Answer
Since we're told p ≠ 1, we know that the antiderivative of

can't be ln x, but is

The integral works out mostly the same way it did in the previous problem:

However, things are a bit different for the final steps. Since p < 1, the exponent p – 1 is negative. This means that as b approaches ∞ the quantity bp – 1 is getting smaller and smaller approaching zero, so the quantity

is getting larger and larger approaching ∞. This integral diverges.
Determine if the statement is true or false. Explain your answer.
If
converges to some finite number L, then

must converge.
Answer
False. If f ( x ) converges to a non-zero number L as x approaches ∞, then the improper integral won't converge.
The integral could converge if L = 0, but even then it wouldn't have to.
Determine if the statement is true or false. Explain your answer.
If
diverges, then

must diverge also.
Answer
True. If the limit of f ( x ) as x approaches ∞ doesn't exist, then that improper integral can't hope to exist either.
Determine if the statement is true or false. Explain your answer.
If
then

must converge.
Answer
False. It's possible to have f ( x ) to converge to 0 as x approaches ∞ but to also have this improper integral diverge. As an example, take
Then
, but

diverges.
Determine if the statement is true or false. Explain your answer.
If
then

must diverge.
Answer
True. If the limit exists but doesn't equal zero, we determined in question 1 that the improper integral diverges. If the limit doesn't exist, we determined in question 2 that the improper integral diverges.
Determine whether the integral converges or diverges. Indicate on a graph the region whose weighted area is given by the integral. If the integral converges, find its value.

Answer
First split up the integral into two improper integrals with only one badly-behaved limit each:

Now work out each of those improper integrals.

Since c is approaching -∞, the quantity ec is approaching 0. So

Now for the other integral.

This limit diverges. Since one of the two improper integrals diverges, the original integral diverges also.
If we had chosen to evaluate the integral

first, we wouldn't have had to bother with the other one at all. If we had only drawn a picture of the function first, we could have seen from the graph that

would diverge! The moral of the story: look at the graph of the function before you start integrating.

If you did look at a graph first and did this the more efficient way, give yourself a pat on the back. Or a carrot.
Determine whether the integral converges or diverges. Indicate on a graph the region whose weighted area is given by the integral. If the integral converges, find its value.

Answer
We have to split up the integral into two improper integrals with only one badly-behaved limit of integration each:

Evaluate the first improper integral:

Looking at a graph of the function arctan x we can see that as b approaches -∞, arctan b approaches
:

This means

Now for the other integral. We look at the graph of arctan x to figure out what arctan b approaches as b approaches ∞.

Putting these results together,

It's a good sign that we got a positive number, since the area between the graph of
and the x-axis all lies on top of the x-axis.