MATH1120 Calculus II

Solution to Supplementary Exercises on Improper Integrals

Alex Fok
November 2, 2013
(1) Solution 1 :
Z ∞ Z b
dx dx
= lim
0 (1 + x )(1 + tan x) b→∞ 0 (1 + x )(1 + tan−1 x)
2 −1 2
Z 1+tan−1 b
du dx
= lim (Let u = 1 + tan−1 x. Then du = )
b→∞ 1 u 1 + x2
= lim ln |1 + tan−1 b|
b→∞
 π
= ln 1 +
2
Therefore the given improper integral converges.
Solution 2 : Note that the improper integral is of type I because there is an infinite
1 1
integration limit. Let f (x) = −1 and g(x) = . Note that 0 ≤
2
(1 + x )(1 + tan x) 1 + x2
f (x) ≤ g(x) and both are continuous for x ∈ [0, ∞). Since
Z ∞ Z b
1
g(x)dx = lim dx
0 b→∞ 0 1 + x2
= lim tan−1 b
b→∞
π
=
2
Z ∞
By the Direct Comparison Test, f (x)dx converges as well.
0
(2) Solution 1 :
Z 1 Z b
4xdx 4xdx
√ = lim √
0 1 − x4 b→1− 0 1 − x4
Z sin−1 b
2 cos θ
= lim dθ ( Let x2 = sin θ. Then 2xdx = cos θdθ)
b→1− 0 | cos θ|
Z sin−1 b p π π
= lim− 2dθ ( 1 − x4 = | cos θ| = cos θ because − ≤ θ ≤ and cos θ ≥ 0 for this range of θ)
b→1 0 2 2
= lim− 2 sin−1 b
b→1
=π

Therefore the given improper integral converges.

Solution 2 : Note that the integral is of type II because the integrand has an Zinfinite
c
du
discontinuity at x = 1. In order to compare with the improper integral of the form p
,
0 u
whose integrand has infinite discontinuity at u = 0, we make the substitution x = 1 − u
1
2

(when x = 1, u = 0).

1 0
−4(1 − u)du
Z Z
4xdx
√ = p
0 1 − x4 1 1 − (1 − u)4
1
4(1 − u)du
Z
= √
0 4u − 6u2 + 4u3 − u4

4(1 − u)
Let f (u) = √ . A comparison function in this type II case can be
4u − 6u2 + 4u3 − u4
obtained by picking the lowest order terms of both the numerator and denominator, as the
4 2
lowest order terms matter the most when u → 0. So g(x) can be chosen to be √ = √ .
4u u
Both f (u) and g(u) are positive and continuous for u ∈ (0, 1],

4(1−u)
√
4u−6u2 +4u3 −u4 2(1 − u)
lim = lim √
u→0 √2 u→0 4 − 6u + 4u2 − u3
u
=1
6= 0

Z 1
2 1
and √ du converges (because it is of type II and p = < 1). By the Limit Comparison
0 u 2
Test, the original integral converges as well.
1
(3) Solution 1 : The given integral is of type I. Let f (x) = √ . A comparison function g(x)
2
x +1
can be obtained by picking the highest order terms of both the numerator and denominator.
1
So g(x) = . Both f (x) and g(x) are positive and continuous for x ∈ [2, ∞),
x

√ 1
x2 +1 x
lim 1 = lim √
x→∞
x
x→∞ x2 + 1
1
= lim q
x→∞
1 + x12
=1
6= 0

Z ∞
1
and dx diverges. By the Limit Comparison test, the original integral diverges as well.
2 x
Solution 2 :
 3

Z ∞ Z b
dx dx
√ = lim √
2 x2 + 1 b→∞ 2 x2 + 1
tan−1 b
sec2 θdθ
Z
= lim
b→∞ tan−1 2 | sec θ|
Z tan−1 b
π π
= lim sec θdθ (| sec θ| = sec θ because − < θ < and sec θ > 0 for θ in that range)
b→∞ tan−1 2 2 2
= lim ln | tan θ + sec θ|
b→∞
p √
= lim ln |b + 1 + b2 | − ln(2 + 5)
b→∞
=∞

Therefore the given integral diverges.

(4) Solution 1 : The integral is of type II because the integrand has an infinite discontinuity
at x = 1. We can apply the trick similar to that used in (2) and make the substitution
x = 1 + u.

Z 2 Z 1
dx du
√ = √
1 x2 − 1 0 2u + u2

1
Let f (u) = √ . g(u) can be chosen by picking the lowest order term of both the
2u + u2
1
numerator and the denominator. So g(u) = √ . Note that both f (u) and g(u) are
2u
positive and continuous for u ∈ (0, 1],

√ 1
√
2 2u
lim 2u+u = lim √
u→0 √1 u→0 2u + u2
2u
1
= lim p u
u→0 1+ 2
=1
6= 0

Z 1
1
and √ du converges. By the Limit Comparison Test, the given improper integral
0 2u
converges as well.
Solution 2 :
4

Z 2 Z 2
dx dx
√ = lim √
1 x2 − 1 a→1+ a x2 − 1
Z sec−1 2
sec θ tan θdθ
= lim+ (Let x = sec θ. Then dx = sec θ tan θdθ)
a→1 sec−1 a | tan θ|
Z sec−1 2
π
= lim sec θdθ (Because a ≤ x ≤ 2 is positive, 0 < θ < , and tan θ > 0)
a→1+ sec−1 a 2
Z sec−1 2
= sec θdθ
sec−1 1
√
= ln(2 + 3)

Therefore the given integral converges.

(5) See Quiz 2 solution.
3x − 1
(6) Solution 1 : The improper integral is of type I. Let f (x) = . g(x) can be chosen to
4x3 − x2
3x 3
be = 2 . Note that both f (x) and g(x) are positive and continuous for x ∈ [1, ∞),
4x3 4x

3x−1
f (x) 4x3 −x2
lim = lim 3
x→∞ g(x) x→∞
4x2
12x − 4x2
3
= lim
x→∞ 12x3 − 3x2
1
1 − 3x
= lim
x→∞ 1 − 1
4x
=1
6= 0

Z ∞
3
and dx converges. By the Limit Comparison Test, the given improper integral
1 4x2
converges as well.
Solution 2 : Resolve the integrand into partial fractions:

3x − 1 3x − 1 A B C
= 2 = + 2+
4x3 − x2 x (4x − 1) x x 4x − 1

3x − 1 = Ax(4x − 1) + B(4x − 1) + Cx2

= (4A + C)x2 + (4B − A)x − B
 5

 
4A + C = 0
 A
 =1
4B − A = 3 =⇒ B =1 .
 
B =1 C = −4
 

Z ∞   Z b  
1 1 4 1 1 4
+ 2− dx = lim + − dx
1 x x 4x − 1 b→∞ 1 x x2 4x − 1
1
= lim (ln b + 1 − − ln(4b − 1) + ln 3)
b→∞ b
b
= lim ln + 1 + ln 3
b→∞ 4b − 1
1
= ln + 1 + ln 3
4
3
= 1 + ln
4
Remark 0.1. While it is legitimate to split the interval of integration, one may not split
the integrand of any improper integral. It is incorrect to conclude that the given improper
integral in this question diverges because
Z ∞  Z ∞ Z ∞ Z ∞
1 1 4 1 1 4
+ 2− dx 6= dx + dx − dx
1 x x 4x − 1 1 x 1 x2 1 4x − 1

The LHS converges while the RHS diverges.

(7) Solution 1 : The given improper integral is of mixed type because it has two infinite inte-
gration limits. So
Z ∞ Z 0 Z ∞
dx dx dx
= +
0 e + e−x
x
−∞ e + e−x
x
0 ex + e−x

1 1
When x → ∞, ex is the dominant term. Let f (x) = , g(x) = x = e−x . Since
ex + e−x e
0 ≤ f (x) ≤ g(x) and both are continuous for x ∈ [0, ∞),
Z ∞ Z ∞
g(x)dx = e−x dx
0 0
Z b
= lim e−x dx
b→∞ 0
−b
= lim (e + 1)
b→∞
=1
Z ∞
by the Direct Comparison Test, f (x)dx converges as well.
0
−x
When x → ∞, e is the dominant term. So we can let the comparison function for the
1
interval of integration (−∞, 0] be h(x) = −x = ex . Since 0 ≤ f (x) ≤ h(x) and both are
e
6

continuous for x ∈ (−∞, 0],

Z 0 Z 0
h(x)dx = lim ex dx
−∞ a→−∞ a
= lim (1 − ea )
a→−∞
=1

Z 0
by the Direct Comparison Test, f (x)dx converges as well. All in all, the given improper
−∞
integral converges.
Solution 2 :

Z Z
dx 1 du du
= · (Let u = ex . Then du = ex dx, dx = )
ex + e−x u + u−1 u u
Z
1
= du
u2 + 1
= tan−1 u + C
= tan−1 ex + C
Z ∞ Z ∞ Z 0
dx dx dx
= +
−∞ ex + e−x 0
x
e +e −x x
−∞ e + e
−x

= lim [tan−1 ex ]b0 + lim [tan−1 ex ]0a

b→∞ a→−∞
π
= − tan−1 1 + tan−1 1 − 0
2
π
=
2

So the given improper integral converges.

1 + sin x
(8) The integral is of type I. Let f (x) = . Note that
x2

1 + sin x 1+1 2
f (x) = 2
≤ 2
= 2
x x x

2
Let g(x) = 2 . Since 0 ≤ f (x) ≤ g(x) and both are continuous for x ∈ [1, ∞), and
Z ∞ x
g(x)dx converges. By the Direct Comparison Test, the given improper integral con-
1
verges.
(9) See Quiz 2 solution.
 7

(10) The integral is of mixed type because it has an infinite integration limit and the integrand
has an infinite discontinuity at x = 1.
Z ∞ Z 2 Z ∞
dx dx dx
= +
1 x ln x 1 x ln x 2 x ln x
Z 2 Z b
dx dx
= lim+ + lim
a→1 a x ln x b→∞ 2 x ln x
Z ln 2 Z ln b
du du dx
= lim+ + lim (Let u = ln x. Then du = )
a→1 ln a u b→∞ ln 2 u x
= lim+ (ln(ln 2) − ln(ln a)) + lim (ln(ln b) − ln(ln 2))
a→1 b→∞
= ln(ln 2) − (−∞) + ∞ − ln(ln 2)
= DNE
So the given improper integral diverges.
1
(11) The integral is of type II with infinite discontinuity at x = 0. Let f (x) = 2 and
x ln x
1
g(x) = . Note that 0 ≥ g(x) ≥ f (x). In order to apply the version of the Direct
x ln x
Comparison Test which requires the non negativity of relevant functions, we shall consider
instead −f (x) and −g(x). We have 0 ≤ −g(x) ≤ −f (x) and both are continuous for
1
x ∈ (0, ]. Note that
2
Z 12 Z 12
dx 1
−g(x)dx = lim − = lim [− ln(ln x)]a2 = ∞
0 a→0+
a x ln x a→0 +

Z 12
By the Direct Comparison Test, −f (x)dx diverges. So does the given improper integral.
0
(12) The integral is of mixed type because there are two infinite integration limits and the
integrand has one infinite discontinuity at x = 0. So
Z ∞ Z −1 Z 0 Z 1 Z ∞
dx dx dx dx dx
2 x
= 2 x
+ 2 x
+ 2 x
+
−∞ x (1 + e ) −∞ x (1 + e ) −1 x (1 + e ) 0 x (1 + e ) 1 x (1 + ex )
2
Z 1
dx 1 1
Consider 2 x
. Let f (x) = 2 x
and g(x) = 2 . Both of them are positive
0 x (1 + e ) x (1 + e ) x
and continuous for x ∈ (0, 1],
1
f (x) x2 (1+ex )
lim = lim+ 1
x→0+ g(x) x→0
x2
1
= lim
x→0+ 1 + ex
1
=
2
6 0
=
Z 1 Z 1
1 dx
and 2
dx diverges. By the Limit Comparison Test, diverges as well, and
0 x 0 x2 (1 + ex )
so does the given improper integral.
8

(13) Applying integration by parts twice, we have

e−x sin x − e−x cos x
Z
e−x cos xdx = +C
2
Z ∞  −x b
e sin x − e−x cos x
e−x cos xdx = lim
0 b→∞ 2 0
 −0
e · sin 0 − e−0 · cos 0

=0−
2
1
=
2
(Here lim e−b sin b = 0 by Squeeze Theorem: −e−b ≤ e−b sin b ≤ e−b and lim ±e−b = 0.
b→∞ b→∞
Same reason for lim e−b cos b = 0) So the given improper integral converges.
b→∞

Remark 0.2. We may not apply the version of the Comparison Tests as in the notes and
the textbook in this case because the integrand e−x cos x is not always nonnegative for all
x ∈ [0, ∞). Nonetheless, we may apply the following
Z Squeeze-Theorem type argument: as
∞ Z ∞
−e−x ≤ e−x cos x ≤ e−x for x ∈ [0, ∞) and both −e−x dx and e−x dx converge, so
0 0
does the given integral.
Z 1
cos t cos t 1
(14) 2
dt is of type II. Let f (t) = 2 and g(t) = 2 . Both f (t) and g(t) are positive
0 t t t
and continuous for t ∈ (0, 1],
f (t)
lim = lim cos t
t→0+ g(t) t→0+
=1
6= 0
Z 1 Z 1
1 cos t
and dt diverges. By the Limit Comparison Test, dt diverges as well. Now
0 t2 0 t2
Z 1 R 1 cos t
cos t x t2
dt ∞
lim x 2
dt = lim 1 (It is indeed of the form )
x→0+ x t x→0 +
x2
∞
− cos
x2
x
= lim+ (By L’Hopital’s Rule)
x→0 − x12
= lim+ cos x
x→0
=1