Improper Integrals

Is the area between the graph of f(x)=1x

and the x-axis over the interval [1,+)

finite or infinite? If this same region is revolved about the x-axis, is the volume finite or infinite? Surprisingly, the area of the region described is infinite, but the volume of the solid obtained by revolving this region about the x-axis is finite.

In this section, we define integrals over an infinite interval as well as integrals of functions containing a discontinuity on the interval. Integrals of these types are called improper integrals. We examine several techniques for evaluating improper integrals, all of which involve taking limits.

Integrating over an Infinite Interval

How should we go about defining an integral of the type a+f(x)dx?

We can integrate atf(x)dx

for any value of t,

so it is reasonable to look at the behavior of this integral as we substitute larger values of t.

[link] shows that atf(x)dx

may be interpreted as area for various values of t.

In other words, we may define an improper integral as a limit, taken as one of the limits of integration increases or decreases without bound.

This figure has three graphs. All the graphs have the same curve, which is f(x). The curve is non-negative, only in the first quadrant, and decreasing. Under all three curves is a shaded region bounded by a on the x-axis an t on the x-axis. The region in the first curve is small, and progressively gets wider under the second and third graph as t moves further to the right away from a on the x-axis.

Definition
  1. Let f(x)

    be continuous over an interval of the form

    [a,+).

    Then


    a+f(x)dx=limt+atf(x)dx,

    provided this limit exists.

  2. Let f(x)

    be continuous over an interval of the form

    (,b].

    Then


    bf(x)dx=limttbf(x)dx,

    provided this limit exists.


    In each case, if the limit exists, then the improper integral is said to converge. If the limit does not exist, then the improper integral is said to diverge.

  3. Let f(x)

    be continuous over

    (,+).

    Then


    +f(x)dx=0f(x)dx+0+f(x)dx,

    provided that

    0f(x)dx

    and

    0+f(x)dx

    both converge. If either of these two integrals diverge, then

    +f(x)dx

    diverges. (It can be shown that, in fact,

    +f(x)dx=af(x)dx+a+f(x)dx

    for any value of

    a.)

In our first example, we return to the question we posed at the start of this section: Is the area between the graph of f(x)=1x

and the x

-axis over the interval [1,+)

finite or infinite?

Finding an Area

Determine whether the area between the graph of f(x)=1x

and the x-axis over the interval [1,+)

is finite or infinite.

We first do a quick sketch of the region in question, as shown in the following graph.

This figure is the graph of the function y = 1/x. It is a decreasing function with a vertical asymptote at the y-axis. In the first quadrant there is a shaded region under the curve bounded by x = 1 and x = 4.

We can see that the area of this region is given by A=11xdx.

Then we have

A=11xdx=limt+1t1xdxRewrite the improper integral as a limit.=limt+ln\|x\|\|1tFind the antiderivative.=limt+(ln\|t\|ln1)Evaluate the antiderivative.=+.Evaluate the limit.

Since the improper integral diverges to +,

the area of the region is infinite.

Finding a Volume

Find the volume of the solid obtained by revolving the region bounded by the graph of f(x)=1x

and the x-axis over the interval [1,+)

about the x

-axis.

The solid is shown in [link]. Using the disk method, we see that the volume V is

V=π1+1x2dx.

This figure is the graph of the function y = 1/x. It is a decreasing function with a vertical asymptote at the y-axis. The graph shows a solid that has been generated by rotating the curve in the first quadrant around the x-axis.

Then we have

V=π1+1x2dx=πlimt+1t1x2dxRewrite as a limit.=πlimt+1x\|1tFind the antiderivative.=πlimt+(1t+1)Evaluate the antiderivative.=π.

The improper integral converges to π.

Therefore, the volume of the solid of revolution is π.

In conclusion, although the area of the region between the x-axis and the graph of f(x)=1/x

over the interval [1,+)

is infinite, the volume of the solid generated by revolving this region about the x-axis is finite. The solid generated is known as Gabriel’s Horn.

Visit this website to read more about Gabriel’s Horn.

Chapter Opener: Traffic Accidents in a City

This is a picture of a city street with a traffic signal. The picture has very busy lanes of traffic in both directions.

In the chapter opener, we stated the following problem: Suppose that at a busy intersection, traffic accidents occur at an average rate of one every three months. After residents complained, changes were made to the traffic lights at the intersection. It has now been eight months since the changes were made and there have been no accidents. Were the changes effective or is the 8-month interval without an accident a result of chance?

Probability theory tells us that if the average time between events is k,

the probability that X,

the time between events, is between a

and b

is given by

P(axb)=abf(x)dxwheref(x)={0ifx<0kekxifx0.

Thus, if accidents are occurring at a rate of one every 3 months, then the probability that X,

the time between accidents, is between a

and b

is given by

P(axb)=abf(x)dxwheref(x)={0ifx<03e−3xifx0.

To answer the question, we must compute P(X8)=8+3e−3xdx

and decide whether it is likely that 8 months could have passed without an accident if there had been no improvement in the traffic situation.

We need to calculate the probability as an improper integral:

P(X8)=8+3e−3xdx=limt+8t3e−3xdx=limt+e−3x\|8t=limt+(e−3t+e−24)3.8×10−11.

The value 3.8×10−11

represents the probability of no accidents in 8 months under the initial conditions. Since this value is very, very small, it is reasonable to conclude the changes were effective.

Evaluating an Improper Integral over an Infinite Interval

Evaluate 01x2+4dx.

State whether the improper integral converges or diverges.

Begin by rewriting 01x2+4dx

as a limit using [link] from the definition. Thus,

01x2+4dx=limxt01x2+4dxRewrite as a limit. =limt12tan−1x2\|t0Find the antiderivative. =12limt(tan−10tan−1t2)Evaluate the antiderivative. =π4.Evaluate the limit and simplify.

The improper integral converges to π4.

Evaluating an Improper Integral on (,+)

Evaluate +xexdx.

State whether the improper integral converges or diverges.

Start by splitting up the integral:

+xexdx=0xexdx+0+xexdx.

If either 0xexdx

or 0+xexdx

diverges, then +xexdx

diverges. Compute each integral separately. For the first integral,

0xexdx=limtt0xexdxRewrite as a limit.=limt(xexex)\|t0Use integration by parts to find theantiderivative. (Hereu=xanddv=ex.)=limt(−1tet+et)Evaluate the antiderivative.=−1.Evaluate the limit. Note:limttetisindeterminate of the form0·.Thus,limttet=limttet=limt−1et=limtet=0byL’Hôpital’s Rule.

The first improper integral converges. For the second integral,

0+xexdx=limt+0txexdxRewrite as a limit.=limt+(xexex)\|0tFind the antiderivative.=limt+(tetet+1)Evaluate the antiderivative.=limt+((t1)et+1)Rewrite.(tetetis indeterminate.)=+.Evaluate the limit.

Thus, 0+xexdx

diverges. Since this integral diverges, +xexdx

diverges as well.

Evaluate −3+exdx.

State whether the improper integral converges or diverges.

e3,

converges

Hint
−3+exdx=limt+−3texdx

Integrating a Discontinuous Integrand

Now let’s examine integrals of functions containing an infinite discontinuity in the interval over which the integration occurs. Consider an integral of the form abf(x)dx,

where f(x)

is continuous over [a,b)

and discontinuous at b.

Since the function f(x)

is continuous over [a,t]

for all values of t

satisfying a<t<b,

the integral atf(x)dx

is defined for all such values of t.

Thus, it makes sense to consider the values of atf(x)dx

as t

approaches b

for a<t<b.

That is, we define abf(x)dx=limtbatf(x)dx,

provided this limit exists. [link] illustrates atf(x)dx

as areas of regions for values of t

approaching b.

This figure has three graphs. All the graphs have the same curve, which is f(x). The curve is non-negative, only in the first quadrant, and increasing. Under all three curves is a shaded region bounded by a on the x-axis an t on the x-axis. There is also a vertical asymptote at x = b. The region in the first curve is small, and progressively gets wider under the second and third graph as t gets further from a, and closer to b on the x-axis.

We use a similar approach to define abf(x)dx,

where f(x)

is continuous over (a,b]

and discontinuous at a.

We now proceed with a formal definition.

Definition
  1. Let f(x)

    be continuous over

    [a,b).

    Then,


    abf(x)dx=limtbatf(x)dx.
  2. Let f(x)

    be continuous over

    (a,b].

    Then,


    abf(x)dx=limta+tbf(x)dx.

    In each case, if the limit exists, then the improper integral is said to converge. If the limit does not exist, then the improper integral is said to diverge.

  3. If f(x)

    is continuous over

    [a,b]

    except at a point

    c

    in

    (a,b),

    then


    abf(x)dx=acf(x)dx+cbf(x)dx,

    provided both

    acf(x)dx

    and

    cbf(x)dx

    converge. If either of these integrals diverges, then

    abf(x)dx

    diverges.

The following examples demonstrate the application of this definition.

Integrating a Discontinuous Integrand

Evaluate 0414xdx,

if possible. State whether the integral converges or diverges.

The function f(x)=14x

is continuous over [0,4)

and discontinuous at 4. Using [link] from the definition, rewrite 0414xdx

as a limit:

0414xdx=limt40t14xdxRewrite as a limit.=limt4(−24x)\|0tFind the antiderivative.=limt4(−24t+4)Evaluate the antiderivative.=4.Evaluate the limit.

The improper integral converges.

Integrating a Discontinuous Integrand

Evaluate 02xlnxdx.

State whether the integral converges or diverges.

Since f(x)=xlnx

is continuous over (0,2]

and is discontinuous at zero, we can rewrite the integral in limit form using [link]:

02xlnxdx=limt0+t2xlnxdxRewrite as a limit.=limt0+(12x2lnx14x2)\|t2Evaluatexlnxdxusing integration by partswithu=lnxanddv=x.=limt0+(2ln2112t2lnt+14t2).Evaluate the antiderivative.=2ln21.Evaluate the limit.limt0+t2lntis indeterminate.To evaluate it, rewrite as a quotient and applyL’Hôpital’s rule.

The improper integral converges.

Integrating a Discontinuous Integrand

Evaluate −111x3dx.

State whether the improper integral converges or diverges.

Since f(x)=1/x3

is discontinuous at zero, using [link], we can write

−111x3dx=−101x3dx+011x3dx.

If either of the two integrals diverges, then the original integral diverges. Begin with −101x3dx:

−101x3dx=limt0−1t1x3dxRewrite as a limit.=limt0(12x2)\|−1tFind the antiderivative.=limt0(12t2+12)Evaluate the antiderivative.=+.Evaluate the limit.

Therefore, −101x3dx

diverges. Since −101x3dx

diverges, −111x3dx

diverges.

Evaluate 021xdx.

State whether the integral converges or diverges.

+,

diverges

Hint

Write 021xdx

in limit form using [link].

A Comparison Theorem

It is not always easy or even possible to evaluate an improper integral directly; however, by comparing it with another carefully chosen integral, it may be possible to determine its convergence or divergence. To see this, consider two continuous functions f(x)

and g(x)

satisfying 0f(x)g(x)

for xa

([link]). In this case, we may view integrals of these functions over intervals of the form [a,t]

as areas, so we have the relationship

0atf(x)dxatg(x)dxforta.

This figure has two graphs. The graphs are f(x) and g(x). The first graph f(x) is a decreasing, non-negative function with a horizontal asymptote at the x-axis. It has a sharper bend in the curve compared to g(x). The graph of g(x) is a decreasing, non-negative function with a horizontal asymptote at the x-axis.

Thus, if

a+f(x)dx=limt+atf(x)dx=+,

then

a+g(x)dx=limt+atg(x)dx=+

as well. That is, if the area of the region between the graph of f(x)

and the x-axis over [a,+)

is infinite, then the area of the region between the graph of g(x)

and the x-axis over [a,+)

is infinite too.

On the other hand, if

a+g(x)dx=limt+atg(x)dx=L

for some real number L,

then

a+f(x)dx=limt+atf(x)dx

must converge to some value less than or equal to L,

since atf(x)dx

increases as t

increases and atf(x)dxL

for all ta.

If the area of the region between the graph of g(x)

and the x-axis over [a,+)

is finite, then the area of the region between the graph of f(x)

and the x-axis over [a,+)

is also finite.

These conclusions are summarized in the following theorem.

A Comparison Theorem

Let f(x)

and g(x)

be continuous over [a,+).

Assume that 0f(x)g(x)

for xa.

  1. If a+f(x)dx=limt+atf(x)dx=+,

    then

    a+g(x)dx=limt+atg(x)dx=+.
  2. If a+g(x)dx=limt+atg(x)dx=L,

    where

    L

    is a real number, then

    a+f(x)dx=limt+atf(x)dx=M

    for some real number

    ML.
Applying the Comparison Theorem

Use a comparison to show that 1+1xexdx

converges.

We can see that

01xex1ex=ex,

so if 1+exdx

converges, then so does 1+1xexdx.

To evaluate 1+exdx,

first rewrite it as a limit:

1+exdx=limt+1texdx=limt+(ex)\|t1=limt+(et+e1)=e1.

Since 1+exdx

converges, so does 1+1xexdx.

Applying the Comparison Theorem

Use the comparison theorem to show that 1+1xpdx

diverges for all p<1.

For p<1,

1/x1/(xp)

over [1,+).

In [link], we showed that 1+1xdx=+.

Therefore, 1+1xpdx

diverges for all p<1.

Use a comparison to show that e+lnxxdx

diverges.

Since e+1xdx=+,

e+lnxxdx

diverges.

Hint
1xlnxx

on [e,+)

Laplace Transforms

In the last few chapters, we have looked at several ways to use integration for solving real-world problems. For this next project, we are going to explore a more advanced application of integration: integral transforms. Specifically, we describe the Laplace transform and some of its properties. The Laplace transform is used in engineering and physics to simplify the computations needed to solve some problems. It takes functions expressed in terms of time and transforms them to functions expressed in terms of frequency. It turns out that, in many cases, the computations needed to solve problems in the frequency domain are much simpler than those required in the time domain.

The Laplace transform is defined in terms of an integral as

L{f(t)}=F(s)=0estf(t)dt.

Note that the input to a Laplace transform is a function of time, f(t),

and the output is a function of frequency, F(s).

Although many real-world examples require the use of complex numbers (involving the imaginary number i=−1),

in this project we limit ourselves to functions of real numbers.

Let’s start with a simple example. Here we calculate the Laplace transform of f(t)=t

. We have

L{t}=0testdt.

This is an improper integral, so we express it in terms of a limit, which gives

L{t}=0testdt=limz0ztestdt.

Now we use integration by parts to evaluate the integral. Note that we are integrating with respect to t, so we treat the variable s as a constant. We have

u=tdv=estdtdu=dtv=1sest.

Then we obtain

limz0ztestdt=limz[[tsest]\|0z+1s0zestdt]=limz[[zsesz+0se−0s]+1s0zestdt]=limz[[zsesz+0]1s[ests]\|0z]=limz[[zsesz]1s2[esz1]]=limz[zsesz]limz[1s2esz]+limz1s2=00+1s2=1s2.
  1. Calculate the Laplace transform of f(t)=1.
  2. Calculate the Laplace transform of f(t)=e−3t.
  3. Calculate the Laplace transform of f(t)=t2.

    (Note, you will have to integrate by parts twice.)


    Laplace transforms are often used to solve differential equations. Differential equations are not covered in detail until later in this book; but, for now, let’s look at the relationship between the Laplace transform of a function and the Laplace transform of its derivative.


    Let’s start with the definition of the Laplace transform. We have


    L{f(t)}=0estf(t)dt=limz0zestf(t)dt.
  4. Use integration by parts to evaluate limz0zestf(t)dt.

    (Let

    u=f(t)

    and

    dv=estdt.)

    After integrating by parts and evaluating the limit, you should see that


    L{f(t)}=f(0)s+1s[L{f(t)}].

    Then,


    L{f(t)}=sL{f(t)}f(0).

    Thus, differentiation in the time domain simplifies to multiplication by s in the frequency domain.


    The final thing we look at in this project is how the Laplace transforms of

    f(t)

    and its antiderivative are related. Let

    g(t)=0tf(u)du.

    Then,


    L{g(t)}=0estg(t)dt=limz0zestg(t)dt.
  5. Use integration by parts to evaluate limz0zestg(t)dt.

    (Let

    u=g(t)

    and

    dv=estdt.

    Note, by the way, that we have defined

    g(t), du=f(t)dt.)

    As you might expect, you should see that


    L{g(t)}=1s·L{f(t)}.

    Integration in the time domain simplifies to division by s in the frequency domain.

Key Concepts

Key Equations

Evaluate the following integrals. If the integral is not convergent, answer “divergent.”

24dx(x3)2

divergent

014+x2dx
0214x2dx
π2
11xlnxdx
1xexdx
2e
xx2+1dx

Without integrating, determine whether the integral 11x3+1dx

converges or diverges by comparing the function f(x)=1x3+1

with g(x)=1x3.

Converges

Without integrating, determine whether the integral 11x+1dx

converges or diverges.

Determine whether the improper integrals converge or diverge. If possible, determine the value of the integrals that converge.

0excosxdx

Converges to 1/2.

1lnxxdx
01lnxxdx

−4

01lnxdx
1x2+1dx
π
15dxx1
−22dx(1+x)2

diverges

0exdx
0sinxdx

diverges

ex1+e2xdx
01dxx3

1.5

02dxx3
−12dxx3

diverges

01dx1x2
031x1dx

diverges

15x3dx
355(x4)2dx

diverges

Determine the convergence of each of the following integrals by comparison with the given integral. If the integral converges, find the number to which it converges.

1dxx2+4x;

compare with 1dxx2.

1dxx+1;

compare with 1dx2x.

Both integrals diverge.

Evaluate the integrals. If the integral diverges, answer “diverges.”

1dxxe
01dxxπ

diverges

01dx1x
01dx1x

diverges

0dxx2+1
−11dx1x2
π
01lnxxdx
0eln(x)dx

0.0

0xexdx
x(x2+1)2dx

0.0

0exdx

Evaluate the improper integrals. Each of these integrals has an infinite discontinuity either at an endpoint or at an interior point of the interval.

09dx9x

6.0

−271dxx2/3
03dx9x2
π2
624dttt236
04xln(4x)dx
8ln(16)4
03x9x2dx

Evaluate .5tdx1x2.

(Be careful!) (Express your answer using three decimal places.)

1.047

Evaluate 14dxx21.

(Express the answer in exact form.)

Evaluate 2dx(x21)3/2.

−1+23

Find the area of the region in the first quadrant between the curve y=e−6x

and the x-axis.

Find the area of the region bounded by the curve y=7x2,

the x-axis, and on the left by x=1.

7.0

Find the area under the curve y=1(x+1)3/2,

bounded on the left by x=3.

Find the area under y=51+x2

in the first quadrant.

5π2

Find the volume of the solid generated by revolving about the x-axis the region under the curve y=3x

from x=1

to x=.

Find the volume of the solid generated by revolving about the y-axis the region under the curve y=6e−2x

in the first quadrant.

3π

Find the volume of the solid generated by revolving about the x-axis the area under the curve y=3ex

in the first quadrant.

The Laplace transform of a continuous function over the interval [0,)

is defined by F(s)=0esxf(x)dx

(see the Student Project). This definition is used to solve some important initial-value problems in differential equations, as discussed later. The domain of F is the set of all real numbers s such that the improper integral converges. Find the Laplace transform F of each of the following functions and give the domain of F.

f(x)=1
1s,s>0
f(x)=x
f(x)=cos(2x)
ss2+4,s>0
f(x)=eax

Use the formula for arc length to show that the circumference of the circle x2+y2=1

is 2π.

Answers will vary.

A function is a probability density function if it satisfies the following definition: f(t)dt=1.

The probability that a random variable x lies between a and b is given by P(axb)=abf(t)dt.

Show that f(x)={0ifx<07e−7xifx0

is a probability density function.

Find the probability that x is between 0 and 0.3. (Use the function defined in the preceding problem.) Use four-place decimal accuracy.

0.8775

Chapter Review Exercises

For the following exercises, determine whether the statement is true or false. Justify your answer with a proof or a counterexample.

exsin(x)dx

cannot be integrated by parts.

1x4+1dx

cannot be integrated using partial fractions.

False

In numerical integration, increasing the number of points decreases the error.

Integration by parts can always yield the integral.

False

For the following exercises, evaluate the integral using the specified method.

x2sin(4x)dx

using integration by parts

1x2x2+16dx

using trigonometric substitution

x2+1616x+C
xln(x)dx

using integration by parts

3xx3+2x25x6dx

using partial fractions

110(4ln(2x)+5ln(x+1)9ln(x+3))+C
x5(4x2+4)5/2dx

using trigonometric substitution

4sin2(x)sin2(x)cos(x)dx

using a table of integrals or a CAS

4sin2(x)sin(x)x2+C

For the following exercises, integrate using whatever method you choose.

sin2(x)cos2(x)dx
x3x2+2dx
115(x2+2)3/2(3x24)+C
3x2+1x42x3x2+2xdx
1x4+4dx
116ln(x2+2x+2x22x+2)18tan−1(1x)+18tan−1(x+1)+C
3+16x4x4dx

For the following exercises, approximate the integrals using the midpoint rule, trapezoidal rule, and Simpson’s rule using four subintervals, rounding to three decimals.

[T] 12x5+2dx

M4=3.312,T4=3.354,S4=3.326

[T] 0πesin(x2)dx

[T] 14ln(1/x)xdx

M4=−0.982,T4=−0.917,S4=−0.952

For the following exercises, evaluate the integrals, if possible.

11xndx,

for what values of n

does this integral converge or diverge?

1exxdx

approximately 0.2194

For the following exercises, consider the gamma function given by Γ(a)=0eyya1dy.

Show that Γ(a)=(a1)Γ(a1).

Extend to show that Γ(a)=(a1)!,

assuming a

is a positive integer.

The fastest car in the world, the Bugati Veyron, can reach a top speed of 408 km/h. The graph represents its velocity.

This figure has a graph in the first quadrant. It increases to where x is approximately 03:00 mm:ss and then drops off steep. The maximum height of the graph, here the drop occurs is approximately 420 km/h.

[T] Use the graph to estimate the velocity every 20 sec and fit to a graph of the form v(t)=aexpbxsin(cx)+d.

(Hint: Consider the time units.)

[T] Using your function from the previous problem, find exactly how far the Bugati Veyron traveled in the 1 min 40 sec included in the graph.

Answers may vary. Ex: 9.405

km

Glossary

improper integral
an integral over an infinite interval or an integral of a function containing an infinite discontinuity on the interval; an improper integral is defined in terms of a limit. The improper integral converges if this limit is a finite real number; otherwise, the improper integral diverges

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 International License.

You can also download for free at http://cnx.org/contents/9a1df55a-b167-4736-b5ad-15d996704270@5.1

Attribution: