Professional Documents
Culture Documents
7.6 Improper Integrals: Preliminary Questions
7.6 Improper Integrals: Preliminary Questions
SOLUTION
(a) The integral is improper because one of the limits of integration is infinite. Because the power of x in the integrand is less than
1, this integral converges.
(b) The integral is improper because the integrand is undefined at x D 0. Because the power of x in the integrand is less than 1,
this integral diverges.
(c) The integral is improper because one of the limits of integration is infinite. Because the power of x in the integrand is greater
than 1, this integral diverges.
(d) The integral is improper because the integrand is undefined at x D 0. Because the power of x in the integrand is greater than
1, this integral converges.
Z =2
2. Is cot x dx an improper integral? Explain.
0
SOLUTION Because the integrand cot x is undefined at x D 0, this is an improper integral.
Z b
1
3. Find a value of b > 0 that makes 2
dx an improper integral.
0 x 4
SOLUTION Any value of b satisfying jbj 2 will make this an improper integral.
Z 1
dx
4. Which comparison would show that converges?
0 x C ex
SOLUTION Note that, for x > 0,
1 1 x
< x De :
x C ex e
Moreover
Z 1
x
e dx
0
converges. Therefore,
Z 1 1
dx
0 x C ex
converges by the comparison test.
Z 1 x
e
5. Explain why it is not possible to draw any conclusions about the convergence of dx by comparing with the integral
Z 1 1 x
dx
.
1 x
SOLUTION For 1 x < 1,
e x 1
< ;
x x
but
Z 1 dx
1 x
diverges. Knowing that an integral is smaller than a divergent integral does not allow us to draw any conclusions using the compar-
ison test.
S E C T I O N 7.6 Improper Integrals 891
Exercises
1. Which of the following integrals is improper? Explain your answer, but do not evaluate the integral.
Z 2 Z 1 Z 1
dx dx
(a) (b) (c) e x dx
0 x 1=3 1 x 0:2 1
Z 1 Z =2 Z 1
(d) e x dx (e) sec x dx (f) sin x dx
0 0 0
Z 1 Z 1 Z 1
dx
(g) sin x dx (h) p (i) ln x dx
0 0 3 x2 1
Z 3
(j) ln x dx
0
SOLUTION
SOLUTION
Z R ˇR
4=3 1=3 ˇ
ˇ 1=3
1
(a) x dx D 3x ˇ D 3R 3.1/ D 3 1 :
1 1 R1=3
Z 1 Z R
4=3 4=3 1
(b) x dx D lim x dx D lim 3 1 D 3.1 0/ D 3:
1 R!1 1 R!1 R1=3
Z 1
2=3
3. Prove that x dx diverges by showing that
1
Z R
2=3
lim x dx D 1
R!1 1
SOLUTION First evaluate the integral on the interval Œ0; R for 0 < R < 3:
Z R
ˇR
dx ˇ
1=2 ˇ 2 2
D 2.3 x/ ˇ D p p :
0 .3 x/3=2 0 3 R 3
Now compute the limit as R ! 3 :
Z 3 Z R
dx dx 2 2
D lim D lim p p D 1I
0 .3 x/3=2 R!3 0 .3 x/3=2 R!3 3 R 3
thus, the integral diverges.
In Exercises 5–40, determine whether the improper integral converges and, if so, evaluate it.
Z 1
dx
5. 19=20
1 x
SOLUTION First evaluate the integral over the finite interval Œ1; R for R > 1:
Z R
ˇR
dx ˇ
D 20x 1=20 ˇˇ D 20R1=20 20:
1 x 19=20 1
SOLUTION First evaluate the integral over the finite interval Œ1; R for R > 1:
Z R ˇR
dx ˇ
1=19 ˇ 19 19
D 19x ˇ D . 19/ D 19 :
1 x 20=19 1 R1=19 R1=19
Now compute the limit as R ! 1:
Z 1 Z R
dx dx 19
D lim D lim 19 D 19 0 D 19:
1 x 20=19 R!1 1 x 20=19 R!1 R1=19
Z 4
7. e 0:0001t dt
1
SOLUTION First evaluate the integral over the finite interval ŒR; 4 for R < 4:
Z ˇ4
4
.0:0001/t e .0:0001/t ˇˇ
e dt D ˇ D 10;000 e 0:0004 e .0:0001/R :
R 0:0001 ˇ
R
Z 5 dx
9.
0 x 20=19
SOLUTION The function x 20=19 is infinite at the endpoint 0, so we’ll first evaluate the integral on the finite interval ŒR; 5 for
0 < R < 5:
Z 5 ˇ5
dx ˇ
1=19 ˇ 1=19 1=19 1 1
D 19x ˇ D 19 5 R D 19 :
R x 20=19 R R1=19 51=19
SOLUTION The function x 19=20 is infinite at the endpoint 0, so we’ll first evaluate the integral on the finite interval ŒR; 5 for
0 < R < 5:
Z 5 ˇ5
dx ˇ
D 20x 1=20 ˇˇ D 20 51=20 R1=20 :
R x 19=20 R
Z R p ˇR p p p
dx ˇ
p D 2 4 xˇ D 2 4 R . 2/ 4 D 4 2 4 R:
0 4 x 0
Z ˇ
R x 2 ˇR 1 1 1 1
3 ˇ
x dx D ˇ D D :
2 2ˇ 2R2 2
2.2 / 8 2R2
2
Z ˇ
R dx .x C 1/ 2 ˇR 1 1 1 1
ˇ
D ˇ D D :
0 .x C 1/3 2 ˇ 2.R C 1/2 2.1/2 2 2.R C 1/2
0
Z ˇ
R e 2x ˇR 1 1
2x ˇ 2R 4 4 2R
e dx D ˇ D e e D e e :
2 2 ˇ 2 2
2
Z ˇ1
1 dx x 0:8 ˇˇ
0:2
D ˇ D 1:25 1 R0:8 :
R x 0:8 ˇ
R
SOLUTION First evaluate the integral on the finite interval Œ4; R for R > 4:
Z ˇ
R e 3x ˇR 1 1
3x ˇ 3R 12 12 3R
e dx D ˇ D e e D e e :
4 3 ˇ 3 3
4
Z ˇR
R
3x e 3x ˇˇ 1 3R
e dx D ˇ D e e 12 :
4 3 ˇ 3
4
Z ˇ0
0
3x e 3x ˇˇ 1 e 3R
e dx D ˇ D :
R 3 ˇ 3 3
R
Z ˇ
2 dx .x 1/ 1 ˇ2 1 1 1
ˇ
D ˇ D D 1:
R .x 1/2 1 ˇ 1 R 1 R 1
R
Z 4 dx
24.
2 .x C 2/1=3
SOLUTION The function .x C 2/ 1=3 is infinite at x D 2, so we’ll first evaluate the integral on the interval ŒR; 4 for 2<
R < 4:
Z ˇ4
4 dx 3 ˇ 3 2=3
D .x C 2/2=3 ˇˇ D 6 .R C 2/2=3 :
R .x C 2/1=3 2 R 2
Next, evaluate the definite integral over the interval ŒR; 6 for R > 3:
Z 6 ˇ6
x 2 ˇ 2 p 2
p dx D .x 3/3=2 C 6.x 3/1=2 ˇˇ D 33=2 C 6 3 .R 3/3=2 6.R 3/1=2
R x 3 3 R 3 3
p 2
D8 3 .R 3/3=2 6.R 3/1=2 :
3
Finally, we compute the limit as R ! 3C :
Z 6 Z 6 p p
x x 2
p dx D lim p dx D lim 8 3 .R 3/3=2 6.R 3/1=2 D 8 3:
3 x 3 R!3C R x 3 R!3C 3
Z 1
29. e x cos x dx
0
SOLUTION First evaluate the indefinite integral using Integration by Parts, with u D e x , v 0 D cos x. Then u0 D e x,
v D sin x, and
Z Z Z
e x cos x dx D e x sin x sin x. e x / dx D e x sin x C e x sin x dx:
Now use Integration by Parts again, with u D e x , v 0 D sin x. Then u0 D e x , v D cos x, and
Z Z
x x x x
e cos x dx D e sin x C e cos x e cos x dx :
R x
Solving this equation for e cos x dx, we find
Z
1
e x cos x dx D e x
.sin x cos x/ C C:
2
Thus,
Z R ˇR
1 ˇ sin R cos R sin 0 cos 0 sin R cos R 1
e x
cos x dx D e x
.sin x cos x/ˇˇ D D C ;
0 2 0 2e R 2 2e R 2
and
Z 1
x sin R cos R 1 1 1
e cos x dx D lim R
C D0C D :
0 R!1 2e 2 2 2
Z 1
2x
30. xe dx
1
SOLUTION First evaluate the indefinite integral using Integration by Parts, with u D x and v 0 D e 2x . Then u0 D 1, v D
1 2x , and
2e
Z Z Z
2x 1 2x 1 2x 1 2x 1 2x
xe dx D xe e dx D e C e dx
2 2 2 2
1 2x 1 2x 1 2x .2x C 1/
D xe e CC D e .2x C 1/ C C D C C:
2 4 4 4e 2x
Therefore,
Z Z ˇ !
1
2x
R
2x .2x C 1/ ˇˇR .2R C 1/ 3
xe dx D lim xe dx D lim D lim C :
1 R! 1 1 R!1 4e 2x ˇ1 R!1 4e 2R 4e 2
Z p
1 e x dx
32. p
0 x
p 1 1=2 dx.
SOLUTION Let u D x, du D 2x Then
Z p Z Z
e x dx p dx p
x
p D2 e p D2 e u du D 2e u C C D 2e x
C C:
x 2 x
p
The function e x =px is infinite, so we first evaluate the integral on ŒR; 1 for 0 < R < 1:
Z p
1 e x dx p ˇ1 p
ˇ
p D 2e x ˇ D 2e 2e R
:
R x R
Thus,
Z R
sin x dx D lim .1 cos R/ D 1 lim cos R:
0 R!1 R!1
This limit does not exist, since the value of cos R oscillates between 1 and 1 as R approaches infinity. Hence the integral does not
converge.
Z =2
36. tan x dx
0
S E C T I O N 7.6 Improper Integrals 899
SOLUTION The function tan x is infinite at x D 2, so we’ll first evaluate the integral on Œ0; R for 0 < R < 2:
Z R ˇR
ˇ
tan x dx D ln j sec xjˇˇ D ln j sec Rj:
0 0
Thus,
Z =2 Z R
tan x dx D lim tan x dx D lim ln j sec Rj D 1:
0 R! 2 0 R! 2
Thus,
Z 1
ln x dx D lim .R 1 R ln R/ D 1 lim R ln R:
0 R!0C R!0C
To compute the limit, rewrite the function as a quotient and apply L’Hôpital’s Rule:
Z 1 1
ln R R
ln x dx D 1 lim 1
D 1 lim 1
D 1 lim . R/ D 1 0 D 1:
0 R!0C R R!0C R!0C
R2
Z 2 dx
38.
1 x ln x
SOLUTION Evaluate the indefinite integral using substitution, with u D ln x, du D .1=x/ dx. Then
Z Z
dx du
D D ln juj C C D ln j ln xj C C:
x ln x u
Thus,
Z 2 dx ˇ2
D ln j ln xjˇR D ln.ln 2/ ln.ln R/;
R x ln x
and
Z 2 dx
D lim ln.ln 2/ ln.ln R/ D ln.ln 2/ lim ln.ln R/ D 1:
1 x ln x R!1 C R!1C
Thus,
Z 1 ln x 1 1 ln R C 1
dx D lim ln R C 1 D 1 C lim D 1:
0 x2 R!0C R R R!0C R
The integral does not converge.
900 CHAPTER 7 TECHNIQUES OF INTEGRATION
Z 1 ln x
40. dx
1 x2
SOLUTION Use Integration by Parts, with u D ln x and v 0 D x 2 . Then u0 D x 1 , v D x 1, and
Z Z
ln x 1 1 1
dx D ln x C x 2 dx D ln x C C:
x2 x x x
Thus,
Z ˇ
R ln x 1 1 ˇˇR 1 1 1 1 1 1
dx D ln x ˇ D ln R ln 1 D1 ln R :
1 x2 x x 1 R R 1 1 R R
Use L’Hôpital’s Rule to compute the limit:
Z 1 1
ln x 1 1 ln R R 0
dx D lim 1 ln R D1 lim 0D1 lim D1 D 1:
1 x2 R!1 R R R!1 R R!1 1 1
Z 1 dx
41. Let I D .
4 .x 2/.x 3/
(a) Show that for R > 4,
Z R ˇ ˇ
dx ˇR 3 ˇˇ 1
D ln ˇˇ ln
4 .x 2/.x 3/ R 2ˇ 2
(b) Then show that I D ln 2.
SOLUTION
(a) The partial fraction decomposition takes the form
1 A B
D C :
.x 2/.x 3/ x 2 x 3
Clearing denominators gives us
1 A B
D C :
x.2x C 5/ x 2x C 5
Clearing denominators gives us
1 D A.2x C 5/ C Bx:
Thus,
ˇ ˇ
1 ˇˇ R ˇˇ 1 1 1 1 1 1 1 7
I D lim ln ˇ ln D ln ln D ln :
R!1 5 2R C 5 ˇ 5 7 5 2 5 7 5 2
Z 1 dx
43. Evaluate I D or state that it diverges.
0 x.2x C 5/
SOLUTION The partial fraction decomposition takes the form
1 A B
D C :
x.2x C 5/ x 2x C 5
Clearing denominators gives us
1 D A.2x C 5/ C Bx:
Thus,
ˇ ˇ
1 1 1 ˇˇ R ˇˇ
I D lim ln ln ˇ D 1:
R!0C 5 7 5 2R C 5 ˇ
The integral does not converge.
Z 1
dx
44. Evaluate I D or state that it diverges.
2 .x C 3/.x C 1/2
SOLUTION The partial fraction decomposition takes the form
1 A B C
2
D C C :
.x C 3/.x C 1/ xC3 xC1 .x C 1/2
Clearing denominators gives us
Thus
ˇ ˇ
1 ˇˇ R C 3 ˇˇ 1 1 5 1 1 1 5
I D lim ln ˇ ln C D ln :
R!1 4 R C 1ˇ 2.R C 1/ 4 3 6 6 4 3
In Exercises 45–48, determine whether the doubly infinite improper integral converges and, if so, evaluate it. Use definition (2).
Z 1
x dx
45. 2
1 1Cx
902 CHAPTER 7 TECHNIQUES OF INTEGRATION
It follows that
Z 1 x dx
1 1 C x2
diverges.
Z 1
jxj
46. e dx
1
SOLUTION First, we find
Z 1 Z 1 Z R
jxj x x R
e dx D e dx D lim e dx D lim 1 e D 1I
0 0 R!1 0 R!1
Z 0 Z 0 Z 0
jxj
e dx D e x dx D lim e x dx D lim 1 e R D 1I
1 1 R! 1 R R! 1
and
Z 1
jxj
e dx D 1 C 1 D 2:
1
Z 1
x2
47. xe dx
1
SOLUTION First note that
Z
x2 1 x2
xe dx D e C C:
2
Thus,
Z 1 Z R
x2 x2 1 1 R21
xe dx D lim xe dx D lim e I D
0 R!1 0 R!1 2 2 2
Z 0 Z 0
x2 x2 1 1 R2 1
xe dx D lim xe dx D lim C e D I
1 R! 1 R R! 1 2 2 2
and
Z 1
x2 1 1
xe dx D D 0:
1 2 2
Z 1 dx
48.
1 .x C 1/3=2
2
SOLUTION First, we evaluate the indefinite integral using the trigonometric substitution x D tan , dx D sec2 d. Then
Z Z Z
dx sec2 x
D d D cos d D sin C C D p C C:
.1 C x 2 /3=2 sec3 1 C x2
Thus,
Z 1 Z R
dx dx R
D lim D lim p D 1I
0 .1 C x 2 /3=2 R!1 0 .1 C x 2 /3=2 R!1 1 C R2
Z 0 Z 0
dx dx R
D lim D lim p D 1I
1 .1 C x 2 /3=2 R! 1 R .1 C x 2 /3=2 R! 1 1 C R2
and
Z 1 dx
D 1 C 1 D 2:
1 .1 C x 2 /.3=2/
S E C T I O N 7.6 Improper Integrals 903
Z 1 Z 0 Z 1
dx dx dx
49. Define J D as the sum of the two improper integrals C . Show that J converges and that
1 x 1=3 1 x 1=3 0 x 1=3
J D 0.
SOLUTION Note that since x 1=3 is an odd function, one might expect this integral over a symmetric interval to be zero. To prove
this, we start by evaluating the indefinite integral:
Z
dx 3
D x 2=3 C C
x 1=3 2
Then
Z Z ˇ
0 dx dx 3 2=3 ˇˇR
R 3 2=3 3 3
D lim D lim x ˇ D lim R D
1 x 1=3 R!0 1 x
1=3 R!0 2 1 R!0 2 2 2
Z 1 Z 1 ˇ1
dx dx 3 2=3 ˇˇ 3 3 2=3 3
D lim D lim x ˇ D lim R D
0 x 1=3 R!0C R x 1=3 R!0C 2 R 2 R!0 C 2 2
so that
Z 1 Z 0 Z 1
dx dx dx 3 3
J D D C D C D0
1 x 1=3 1 x 1=3 0 x 1=3 2 2
Z 1 dx
50. Determine whether J D (defined as in Exercise 49) converges.
1 x2
SOLUTION We have
Z
dx 1
D CC
x2 x
so that
Z Z ˇ !
0 dx R dx 1 ˇˇR 1 1
D lim D lim ˇ D lim C 1 D1 lim D1
1 x2 R!0 1 x2 R!0 x 1 R!0 R R!0 R
Z Z ˇ !
1 dx 1 dx 1 ˇˇ1 1 1
D lim D lim D lim 1 C D 1 C lim D1
0 x2 R!0C R x2 R!0C x ˇR R!0C R R!0C R
If p D 1, then
Z 1 Z 1 ˇ1
dx dx ˇ
D D ln x ˇˇ D ln 1 ln R D ln RI and
R xp R x R
Z 1 dx
D lim . ln R/ D 1:
0 x R!0C
1
0.8
0.6
0.4
0.2
x
−4 −2 2 4
Since .1 C x 2 / 1 is an even function, we can first compute the area under the graph for x > 0:
Z R ˇR
dx 1 ˇ
2
D tan x ˇ D tan 1 R tan 1 0 D tan 1 R:
0 1Cx 0
Thus,
Z 1 dx 1
D lim tan RD :
0 1 C x2 R!1 2
By symmetry, we have
Z 1 Z 0 Z 1
dx dx dx
D D C D :
C
1 1 C x2 1 1 C x21 C x2
2 2 0
Z 1
1 1 dx
54. Show that p 2 for all x, and use this to prove that p converges.
x4 C 1 x 1 x4 C 1
p p
SOLUTION Since x 4 C 1 x 4 D x 2 , it follows that
1 1
p 2:
x4 C 1 x
The integral
Z 1 dx
1 x2
converges by Theorem 2, since 2 > 1. Therefore, by the comparison test,
Z 1
dx
p converges:
1 x4 C 1
Z 1 Z 1
dx
55. Show that converges by comparing with x 3 dx.
1 x3 C 4 1
Z 1
SOLUTION The integral x 3 dx converges because 3 > 1. Since x 3 C 4 x 3 , it follows that
1
1 1
3:
x3 C 4 x
Therefore, by the comparison test,
Z 1 dx
converges:
1 x3 C 4
S E C T I O N 7.6 Improper Integrals 905
Z 1
dx
56.Z Show that converges by comparing with
1 2 x3 4
3
2x dx.
2
Z 1 Z 1
3 3
SOLUTION The integral x dx converges because 3 > 1. If x dx D M < 1, then
1 1
Z 1 Z 1
3 3
2x dx D 2 x dx D 2M
1 1
1 2
:
x3 4 x3
Therefore, by the comparison test:
Z 1 dx
converges:
2 x3 4
x2 x
R1 x2
57. Show that 0 e e for x 1 (Figure 1). Use the Comparison Test to show that 0 e dx converges.
Hint: It suffices (why?) to make the comparison for x 1 because
Z 1 Z 1 Z 1
2 2 2
e x dx D e x dx C e x dx
0 0 1
y
1 y = e −x
2
y = e −|x|
x
−4 −3 −2 −1 1 2 3 4
2
FIGURE 1 Comparison of y D e jxj and y D e x .
2
SOLUTION For x 1, x 2 x, so x 2 x and e x e x . Now
Z 1 Z 1
x2
e x dx converges; so e dx converges
1 1
both converge, so
Z 1 Z 1
x2 x2
e dx and e dx
1 1
2
must also converge by the comparison test. Because e x is continuous on Œ 1; 1, it follows that
Z 1 Z 1 Z 1 Z 1
2 2 2 2
e x dx D e x dx C e x dx C e x dx
1 1 1 1
converges.
906 CHAPTER 7 TECHNIQUES OF INTEGRATION
Z 1
sin x 1
59. Show that dx converges.
x2 1
Z 1
1 sin x 2
SOLUTION Let f .x/ D . Since f .x/ and 2x 2 dx D 2, it follows that
x2 x2 1
Z 1
1 sin x
dx converges
1 x2
by the comparison test.
xa
60. Let a > 0. Recall that lim D 1 (by Exercise 64 in Section 4.5).
x!1 ln x
a
(a) Show that x > 2 ln x for all x sufficiently large.
a
e x < x 2 for all x sufficiently large.
(b) Show that Z
1
xa
(c) Show that e dx converges.
1
SOLUTION
(a) Since lim x a = ln x D 1, there must be some number M > 0 such that, for all x > M ,
x!1
xa
> 2:
ln x
But this means that, for all x > M ,
x a > 2 ln x:
so that
xa 2
e < e ln x Dx 2
:
(c) By the above calculations, we can use the comparison test on the interval ŒM; 1/:
Z 1 Z 1
dx a
2
converges ) e x dx also converges:
M x M
In Exercises 61–74, use the Comparison Test to determine whether or not the integral converges.
Z 1
1
61. p dx
1 5
x C2
p p
SOLUTION Since x 5 C 2 x 5 D x 5=2 , it follows that
1 1
p :
x5 C 2 x 5=2
Z 1
The integral dx=x 5=2 converges because 5
2 > 1. Therefore, by the comparison test:
1
Z 1 dx
p also converges:
1 x5 C 2
Z 1 dx
62.
.x 3 C 2x C 4/1=2
1
p p
SOLUTION For all x 1, x 3 C 2x C 4 x 3 D x 3=2 . Thus
1 1
p :
x3 C 2x C 4 x 3=2
Z 1
The integral dx=x 3=2 converges because 3
2 > 1. Therefore, by the comparison test,
1
Z 1 dx
p also converges:
1 x 3 C 2x C 4
S E C T I O N 7.6 Improper Integrals 907
Z 1 dx
63. p
3 x 1
p p
SOLUTION Since x x 1, we have (for x > 1)
1 1
p p :
x x 1
Z 1 Z 1
p
The integral dx= x D dx=x 1=2 diverges because 1
2 < 1. Since the function x 1=2 is continuous (and therefore finite)
1 Z 1
1
on Œ1; 3, we also know that dx=x 1=2 diverges. Therefore, by the comparison test,
3
Z 1 dx
p also diverges:
3 x 1
Z 5 dx
64.
0 x 1=3 C x3
SOLUTION For 0 x 5, x 1=3 C x 3 x 1=3 , so that
1 1
1=3 :
x 1=3 Cx 3 x
Z 5
1=3
The integral x dx converges; therefore, by the comparison test
0
Z 5 dx
also converges.
0 x 1=3 C x3
Z 1
.xCx 1/
65. e dx
1
1 1
SOLUTION For all x 1, x > 0 so x C x x. Then
1 .xCx 1/ x
xCx x and e e :
Z 1
x
The integral e dx converges by direct computation:
1
Z 1 Z R ˇR
x x
ˇ
xˇ R 1 1 1
e dx D lim e dx D lim e ˇ D lim e Ce D0Ce De :
1 R!1 1 R!1 1 R!1
j sin xj 1
p p :
x x
The integral
Z 1 Z 1
dx dx
p D
0 x 0 x 1=2
1
converges, since 2 < 1. Therefore, by the comparison test,
Z 1 j sin xj
p dx also converges:
0 x
Z 1 ex
67. dx
0 x2
908 CHAPTER 7 TECHNIQUES OF INTEGRATION
1 ex
< :
x2 x2
Z 1
The integral dx=x 2 diverges since 2 > 1. Therefore, by the comparison test,
0
Z 1 ex
also diverges:
0 x2
Z 1 1
68. dx
1 x 4 C ex
SOLUTION For x > 1, x 4 C e x x 4 , and
1 1
4:
x 4 C ex x
Z 1
The integral dx=x 4 converges, since 4 > 1. Therefore, by the comparison test,
0
Z 1 dx
also converges:
1 x 4 C ex
Z 1 1
69. p dx
0 x4 C x
p p
SOLUTION For 0 < x < 1, x 4 C x x, and
1 1
p p :
x4 C x x
Z 1 p 1
The integral .1= x/ dx converges, since p D 2 < 1. Therefore, by the comparison test,
0
Z 1 dx
p also converges:
0 x4 C x
Z 1 ln x
70. dx
1 sinh x
SOLUTION For x > 1, e x < 12 e x , so
ex e x 1 x
sinh x D e :
2 4
Similarly, ln x < x for all x > 1, so
ln x 4x
x for all x 1:
sinh x e
Because
Z 1 ˇ1 Z 1
x
ˇ
xˇ x 8
4xe dx D 4xe ˇ C 4e dx D ;
1 1 1 e
it follows by the comparison test that
Z 1 ln x
dx converges:
1 sinh x
Z 1 dx
71. p
0 x 1=3 C x 3
SOLUTION Note that
Z 1 Z 1 Z 1
dx dx dx
p D p C p
0 x 1=3 C x 3 0 x 1=3 C x 3 1 x 1=3 C x 3
p p
For the first integral, for x 0; x 1=3 C x 3 x 1=3 D x 1=6 ; so that
1 1
p 1=6
x 1=3 C x 3 x
S E C T I O N 7.6 Improper Integrals 909
The integral
Z 1
1=6
x dx
0
xe x C x 2 D x.e x C x/I examining the first integral, for 0 x 1; e x e 1 D e and x 1, so that x.e x C x/ x.e C 1/: It
follows that
Z 1 Z 1 Z 1
1 1 1 1
x 2
dx dx D dx
0 xe C x 0 x.e C 1/ eC1 0 x
Z 1
1
But dx diverges since p D 1. Therefore, by the comparison test,
0 x
Z 1
1
x C x2
dx diverges as well.
0 xe
Thefore the original integral must diverge.
Hint for Exercise 73: Show that for x 1,
1 1
.x C x 2 /1=3 21=3 x 2=3
77. An investment pays a dividend of $250/year continuously forever. If the interest rate is 7%, what is the present value of the
entire income stream generated by the investment?
SOLUTION The present value of the income stream after T years is
Z ˇT
T
0:07t 250e 0:07t ˇˇ 250 0:07T
250 0:07T
250e dt D ˇ D e 1 D 1 e :
0 0:07 ˇ 0:07 0:07
0
79. Compute the present value of an investment that generates income at a rate of 5000te 0:01t dollars per year forever, assuming
an interest rate of 6%.
SOLUTION The present value of the income stream after T years is
Z T Z T
5000te 0:01t e 0:06t
dt D 5000 te 0:05t
dt
0 0
Compute the indefinite integral using Integration by Parts, with u D t and v 0 D e 0:05t . Then u0 D 1, v D . 1=0:05/e 0:05t , and
Z Z
t 1 20
te 0:05t dt D e 0:05t C e 0:05t dt D 20te 0:05t C e 0:05t C C
0:05 0:05 0:05
0:05t
De . 20t 400/ C C:
Thus,
Z T ˇT
5000 te 0:05t
dt D 5000e 0:05t
. 20t 400/ˇ0 D 5000e 0:05T
. 20T 400/ 5000. 400/
0
0:05T
D 2;000;000 5000e .20T C 400/:
81. The solid S obtained by rotating the region below the graph of y D x 1 about the x-axis for 1 x < 1 is called Gabriel’s
Horn (Figure 2).
(a) Use the Disk Method (Section 6.3) to compute the volume of S. Note that the volume is finite even though S is an infinite
region.
(b) It can be shown that the surface area of S is
Z 1 p
A D 2 x 1 1 C x 4 dx
1
Show that A is infinite. If S were a container, you could fill its interior with a finite amount of paint, but you could not paint its
surface with a finite amount of paint.
y = x −1
FIGURE 2
SOLUTION
(a) The volume is given by
Z 1 2
1
V D dx:
1 x
First compute the volume over a finite interval:
Z 2 Z R ˇ
R 1 x 1 ˇR 1 1 1
2 ˇ
dx D x dx D ˇ D D 1 :
1 x 1 1ˇ R 1 R
1
Thus,
Z 1
2 1
V D lim x dx D lim 1 D :
R!1 1 R!1 R
(b) For x > 1, we have
r s p p
1 1 1 x4 C 1 x4 C 1 x4 x2 1
1C 4 D D D 3 D :
x x x x4 x 3 x 3 x x
Z 1 1
The integral dx diverges, since p D 1 1. Therefore, by the comparison test,
1 x
Z 1 r
1 1
1 C 4 dx also diverges:
1 x x
Finally,
Z r
1 1 1
A D 2 1C dx
1 x x4
diverges.
82. Compute the volume of the solid obtained by rotating the region below the graph of y D e jxj=2 about the x-axis for 1<
x < 1.
S E C T I O N 7.6 Improper Integrals 913
SOLUTION The graph of y is symmetric around the y-axis, so it suffices to compute the volume for 0 x 1, where we have
yDe x=2 . Using the disk method,
Z 1 2 Z 1 Z R
V D2 e x=2 dx D 2 e x dx D 2 lim e x dx
0 0 R!1 0
ˇR
ˇ
xˇ R
D lim 2e ˇ D 2 lim .e 1/ D 2
R!1 R!1 0
Therefore V D 2.
83. When a capacitor of capacitance C is charged by a source of voltage V , the power expended at time t is
V2 t =RC 2t =RC
P .t/ D .e e /
R
where R is the resistance in the circuit. The total energy stored in the capacitor is
Z 1
W D P .t/ dt
0
1 2
Show that W D 2CV .
SOLUTION The total energy contained after the capacitor is fully charged is
Z 1
V2 t =RC 2t =RC
W D e e dt:
R 0
which is infinite.
1
Now, suppose p ¤ 1. Using the substitution u D ln x, so that du D x dx, the integral becomes
Z 1=2 Z xD1=2 Z xD1=2 ˇxD1=2
dx du p 1 ˇ
pC1 ˇ
D D u du D u ˇ
R x.ln x/p xDR up p 1
xDR xDR
ˇ1=2
1 ˇ 1 1
D .ln x/ pC1 ˇˇ D .ln.1=2// pC1 .ln.R// pC1
:
p 1 R p 1 p 1
By definition,
Z 1=2 Z 1=2
dx dx 1 pC1 1 pC1
D lim D lim .ln.1=2// .ln R/ :
0 x.ln x/p R!0C R x.ln x/p R!0C p 1 p 1
pC1 1
If p > 1, lim .ln R/ D lim p 1 D 0. If p < 1, lim .ln R/1 p D 1. Therefore, the integral diverges if p < 1 or
R!0C R!0 .ln R/ R!0C
p D 1, and converges if p > 1.
914 CHAPTER 7 TECHNIQUES OF INTEGRATION
85. Conservation of Energy can be used to show that when a mass m oscillates at the end of a spring with spring constant k, the
period of oscillation is
Z p
p 2E=k dx
T D4 m p
0 2E kx 2
p
where E is the total energy of the mass. Show that this is an improper integral with value T D 2 m=k.
p
SOLUTION The integrand is infinite at the upper limit of integration, x D 2E=k, so the integral is improper. Now, let
Z R Z R
p dx p 1 dx
T .R/ D 4 m p D 4 mp q
0 2E kx 2 2E 0 1 k
. 2E /x 2
r r r ! r !
m 2E 1 k p 1 k
D4 sin R D 4 m=k sin R :
2E k 2E 2E
Therefore
r r
m 1 m
T D lim
p T .R/ D 4 sin .1/ D 2 :
R! 2E=k k k
In Exercises 86–89, the Laplace transform of a function f .x/ is the function Lf .s/ of the variable s defined by the improper
integral (if it converges):
Z 1
Lf .s/ D f .x/e sx dx
0
and
Z Z
1 sx ˛
e sx sin ˛x dx D e sin ˛x C e sx cos ˛x dx:
s s
Use Integration by Parts again, with u D cos ˛x, v 0 D e sx . Then u0 D ˛ sin ˛x, v D 1s e sx , and
Z Z
1 sx ˛
e sx cos ˛x dx D e cos ˛x e sx sin ˛x dx:
s s
R
Substituting this into the first equation and solving for e sx sin ˛x dx, we get
Z Z
1 sx ˛ sx ˛2
e sx sin ˛x dx D e sin ˛x e cos ˛x e sx sin ˛x dx
s s2 s2
Z e sx 1s sin ˛x C s˛2 cos ˛x e sx .s sin ˛x C ˛ cos ˛x/
e sx sin ˛x dx D D
1C ˛
2 s2 C ˛2
s2
Thus,
Z R
sx 1 s sin ˛R C ˛ cos ˛R 0C˛ 1 s sin ˛R C ˛ cos ˛R
e sin ˛x dx D D 2 ˛ :
0 s2 C ˛2 e sR 1 s C ˛2 e sR
Finally we take the limit, noting the fact that, for all values of R, js sin ˛R C ˛ cos ˛Rj s C j˛j
1 s sin ˛R C ˛ cos ˛R 1 ˛
Lf .s/ D lim 2 ˛ D 2 .˛ 0/ D 2 :
R!1 s C ˛ 2 e sR s C ˛2 s C ˛2
S E C T I O N 7.6 Improper Integrals 915
First evaluate the indefinite integral using Integration by Parts, with u D cos ˛x and v 0 e sx . Then u0 D ˛ sin ˛x, v D
1 sx , and
se
Z Z
1 sx ˛
e sx cos ˛x dx D e cos ˛x e sx sin ˛x dx:
s s
Use Integration by Parts again, with u D sin ˛x dx and v 0 D e sx . Then u0 D ˛ cos ˛x, v D 1s e sx , and
Z Z
1 sx ˛
e sx sin ˛x dx D e sin ˛x C e sx cos ˛x dx:
s s
R
Substituting this into the first equation and solving for e sx cos ˛x dx, we get
Z Z
1 sx ˛ 1 sx ˛
e sx cos ˛x dx D e cos ˛x e sin ˛x C e sx cos ˛ dx
s s s s
2 Z
1 sx ˛ ˛
D e cos ˛x C 2 e sx sin ˛x e sx cos ˛x dx
s s s2
Z e sx s˛2 sin ˛x 1s cos ˛x e sx .˛ sin ˛x s cos ˛x/
e sx cos ˛x dx D D
1C ˛2 s2 C ˛2
s2
Thus,
Z R
sx 1 ˛ sin ˛R s cos ˛R 0 s
e cos ˛x dx D :
0 s2 C ˛2 e sR 1
Finally we take the limit, noting the fact that, for all values of R, j˛ sin ˛R s cos ˛Rj j˛j C s
1 ˛ sin ˛R s cos ˛R 1 s
Lf .s/ D lim 2 2
s C D 2 2
.s C 0/ D 2 :
R!1 s C ˛ e sR s C˛ s C ˛2
90. When a radioactive substance decays, the fraction of atoms present at time t is f .t/ D e kt , where k > 0 is the
R1 0
decay constant. It Rcan be shown that the average life of an atom (until it decays) is A D 0 tf .t/ dt. Use Integration by Parts
1
to show that A D 0 f .t/ dt and compute A. What is the average decay time of radon-222, whose half-life is 3.825 days?
SOLUTION Let u D t, v 0 D f 0 .t/. Then u0 D 1, v D f .t/, and
Z 1 ˇ1 Z 1
ˇ
AD tf .t/ dt D tf .t/ˇˇ C
0
f .t/ dt:
0 0 0
Now,
Z ˇR
R
kt 1 ˇ
kt ˇ 1 kR
1 kR
e dt D e ˇ D e 1 D 1 e ;
0 k 0 k k
so
1 kR
1 1
A D lim 1 e D .1 0/ D :
R!1 k k k
Because k has units of (time) 1 , A does in fact have the appropriate units of time. To find the average decay time of Radon-222,
we need to determine the decay constant k, given the half-life of 3.825 days. Recall that
ln 2
kD
tn
where tn is the half-life. Thus,
1 tn 3:825
AD D D 5:518 days:
k ln 2 ln 2
Z 1
91. Let Jn D xn e ˛x
dx, where n 1 is an integer and ˛ > 0. Prove that
0
n
Jn D Jn 1
˛
and J0 D 1=˛. Use this to compute J4 . Show that Jn D nŠ=˛ nC1 .
1
SOLUTION Using Integration by Parts, with u D x n and v 0 D e ˛x , we get u0 D nx n 1 , v D ˛e
˛x , and
Z Z
1 n ˛x n
x n e ˛x dx D x e C x n 1 e ˛x dx:
˛ ˛
Thus,
Z ˇR Z 1
1 1 n ˇ
ˇ Cn Rn n
Jn D xne ˛x
dx D lim x e ˛x
ˇ xn 1
e ˛x
dx D lim C 0 C Jn 1:
0 R!1 ˛ 0 ˛ 0 R!1 ˛e ˛R ˛
Use L’Hôpital’s Rule repeatedly to compute the limit:
then we have
n n .n 1/Š nŠ
Jn D Jn 1 D D nC1 :
˛ ˛ ˛n ˛
xn
92. Let a > 0 and n > 1. Define f .x/ D for x ¤ 0 and f .0/ D 0.
e ax
1
(a) Use L’Hôpital’s
R1 Rule to show that f .x/ is continuous at x D 0.
(b) Show that 0 f .x/ dx converges. Hint: Show that f .x/ 2x n e ax if x is large enough. Then use the Comparison Test and
Exercise 91.
SOLUTION
(a) Using L’Hôpital’s Rule, we find
xn nx n 1 0
lim D lim D D 0I
x!0 e ˛x 1 x!0 ˛e ˛x ˛
thus,
lim f .x/ D f .0/;
x!0
SOLUTION
(a) If p D 1, then
Z 1 Z 1
1 ln x
I D x ln x dx D dx:
0 0 x
Let u D ln x, du D .1=x/ dx. Then
Z Z
ln x u2 1
dx D u du D C C D .ln x/2 C C:
x 2 2
Thus,
Z 1 ln x 1 1 1
dx D .ln 1/2 .ln R/2 D .ln R/2 ;
R x 2 2 2
and
1
I D lim .ln R/2 D 1:
R!0C 2
The integral diverges for p D 1.
(b) If p ¤ 1, then use Integration by Parts, with u D ln x and v 0 D x p . Then u0 D 1=x, v D x pC1 =p C 1, and
Z
x pC1 1
Z 1 x pC1 1
Z
x p ln x dx D ln x x pC1 dx D ln x x p dx
pC1 pC1 x pC1 pC1
!
x pC1 1 x pC1 x pC1 1
D ln x CC D ln x C C:
pC1 pC1 pC1 pC1 pC1
The integral diverges for p < 1. On the other hand, if p > 1, then p C 1 > 0, and
1 1 1 1 1
I D C lim RpC1 ln R C lim RpC1 D C0 D :
.p C 1/2 p C 1 R!0C .p C 1/2 R!0C .p C 1/2 .p C 1/2
95. Let
Z x dt x
F .x/ D and G.x/ D
2 ln t ln x
F .x/
Verify that L’Hôpital’s Rule applies to the limit L D lim and evaluate L.
x!1 G.x/
S E C T I O N 7.6 Improper Integrals 919
1
SOLUTION Because ln t < t for t > 2, we have > 1t for t > 2, and so
ln t
Z x Z x
dt dt
F .x/ D > D ln x ln 2
2 ln t 2 t
Thus, F .x/ ! 1 as x ! 1. Moreover, by L’Hôpital’s Rule
1
lim G.x/ D lim D lim x D 1:
x!1 x!1 1=x x!1
F .x/
Thus, lim is of the form 1=1, and L’Hôpital’s Rule applies. Finally,
x!1 G.x/
1
F .x/ ln x ln x 1
L D lim D lim ln x 1
D lim D lim D 1:
x!1 G.x/ x!1 x!1 ln x 1 x!1 1 .1= ln x/
.ln x/2
R1 R1
In Exercises 96–98, an improper integral I D a f .x/ dx is called absolutely convergent if a jf .x/j dx converges. It can be
shown that if I is absolutely convergent, then it is convergent.
Z 1
sin x
96. Show that dx is absolutely convergent.
1 x2
SOLUTION For all x, j sin xj 1. This implies
ˇ ˇ
ˇ sin x ˇ j sin xj 1
ˇ ˇ
ˇ x2 ˇ D x2 x2 :
Z 1
2
The integral x dx converges because p D 2 > 1. Therefore, by the comparison test,
1
Z 1 ˇˇ sin x ˇˇ
ˇ ˇ
ˇ x 2 ˇ dx also converges:
1
It is known that I D 2. However, I is not absolutely convergent. The convergence depends on cancellation, as shown in Figure 3.
y
1
y= x
1
sin x
y= x
x
1 2 3 7
−1
1
y=−x
R1
FIGURE 3 Convergence of 1 .sin x=x/ dx is due to the cancellation arising from the periodic change of sign.
SOLUTION
1
(a) Use Integration by Parts, with u D x and v 0 D sin x. Then u0 D 1=x 2 , v D cos x, and we have
Z Z
R sin x cos x ˇˇR R cos x
dx D ˇ dx:
1 x x 1 1 x2
(b) For all x, j cos xj 1, and therefore
ˇ cos x ˇ j cos xj 1
ˇ ˇ
ˇ 2 ˇD 2
2:
x x x
Z 1
2
The integral x dx converges, because p D 2 > 1. Therefore, by the comparison test,
1
Z 1 ˇ cos x ˇ
ˇ ˇ
ˇ ˇ dx also converges:
1 x2
Z 1
Because .cos x=x 2 / dx converges absolutely, it also converges. By this result,
1
Z " Z # Z
R sin x cos R cos 1 R cos x cos 1 1 cos x
lim dx D lim C dx D 0 C dx D cos 1 M;
R!1 1 x R!1 R 1 1 x2 1 0 x2
Z 1 Z 1
where M D .cos x=x 2 / dx, the existence of which was shown in the argument above. Therefore the integral .sin x=x/ dx
1 1
converges to a finite value.
(c) The integral can be split up as follows:
Z 1 Z 1 Z 1
sin x sin x sin x
dx D dx C dx:
0 x 0 x 1 x
The second integral converges by part (b). For the first integral, if we define f .0/ D 1, then the integrand is continuous on Œ0; 1,
and therefore
Z 1
sin x
dx D N
0 x
where N is some finite value. Thus, we have shown that I converges.
99. The gamma function, which plays an important role in advanced applications, is defined for n 1 by
Z 1
.n/ D t n 1 e t dt
0
(a) Show that the integral defining .n/ converges for n 1 (it actually converges for all n > 0). Hint: Show that t n 1 e t < t 2
for t sufficiently large.
(b) Show that .n C 1/ D n.n/ using Integration by Parts.
(c) Show that .n C 1/ D nŠ if n 1 is an integer. Hint: Use (a) repeatedly. Thus, .n/ provides a way of defining n-factorial
when n is not an integer.
S E C T I O N 7.6 Improper Integrals 921
SOLUTION
(a) By repeated use of L’Hôpital’s Rule, we can compute the following limit:
et et et
lim D lim n
D D lim D 1:
t !1 t nC1 t !1 .n C 1/t t !1 .n C 1/Š
This implies that, for t sufficiently large, we have
e t t nC1 I
therefore
et t nC1
D t2 or tn 1
e t
t 2
:
tn 1 tn 1
Z 1
2
The integral t dt converges because p D 2 > 1. Therefore, by the comparison test,
1
Z 1
tn 1
e t
dt also converges;
M
where M is the value above which the above comparisons hold. Finally, because the function t n 1e t is continuous for all t, we
know that
Z 1
.n/ D t n 1 e t dt converges for all n 1:
0
If n D 1, then
Z 1 ˇR
t
ˇ
tˇ R
.1/ D e dt D lim e ˇ D lim 1 e D 1:
0 R!1 0 R!1
Thus
.n C 1/ D nŠ .1/ D nŠ
nŠ
100. Use the results of Exercise 99 to show that the Laplace transform (see Exercises 86–89 above) of x n is .
s nC1
SOLUTION If f .x/ D x n , then the Laplace transform of f .x/ is
Z 1
Lf .s/ D xne sx
dx
0