Ch 01: Improper Integrals of 1st and 2nd Kinds

Course Title: Real Analysis II Course Code: MTH322

Course instructor: Dr. Atiq ur Rehman Class: MSc-II
Course URL: www.mathcity.org/atiq/sp19-mth322

“The objective of this chapter is to learn about different kind of improper

integral. To learn the meaning of convergence and divergence of improper
integrals. Learn the theory to develop different tests and techniques to find
convergence or divergence of improper integrals”

We discussed Riemann’s integrals of the form  a

b
f ( x) dx under the
restrictions that both f is defined and bounded on a finite interval [ a , b ] . To
extend the concept, we shall relax some condition on definite integral like f on
finite interval or boundedness of f on finite interval.
First of all we recall few things about symbol + ∞ (or ∞ ) and −∞ . These symbols
don’t behave like usual real numbers. Historically, with these concepts,
mathematicians were never very comfortable, and these were some sources of
much confusion and debate. Indeed, the concepts of infinite sets and infinity took
nearly a century for a definite implication. To get the idea of infinity one can read
the articles at following URLs:
 https://www.mathsisfun.com/numbers/infinity.html
 https://en.wikipedia.org/wiki/Infinity
An integral is said to be improper integral if either the function f is unbounded
on [ a, b] or the interval of integration is unbounded. Now we are going to give
formal definitions of improper integrals.
 Definition
b b
The integral  a
f ( x) dx or  a
f dx is called an improper integral of first kind
if a = − ∞ or b = ∞ or both i.e. one or both integration limits are infinite.

 Definition
b
The integral  a
f dx is called an improper integral of second kind if f ( x) is
unbounded with infinite discontinuity at one or more points of a ≤ x ≤ b .
 Remark:
Some time we deal with an improper integral which involves both kinds of
integral at once. It is known as improper integral of mixed kind. It can be break
in to the sum of improper integrals of first and second kinds.
2 Ch 01: Improper Integrals of 1st and 2nd Kinds

 Examples
∞ 1 ∞
1 1
•  2
dx ,  dx and  ( x 2 + 1)dx are
0
1+ x −∞
x−2 −∞
examples of improper integrals of first kind.

1 1
1 1
•  dx and  dx are examples of
−1
x 0
2x −1
improper integrals of second kind.

∞
1
• 0 x dx is an example of improper integral of
mixed kind and it can be written as follow:
∞ 1 ∞
1 1 1
0 x dx = 0 x dx + 1 x dx

 MCQs
(i) Which of the following is an improper integral of 1st kind.
2 ∞ ∞
1 1
(I)  dx (II)  2 dx (III)  (2t + 1)dt
1
x 1
x −∞
A. I and III only B. III only
C. II only D. II and III only
(ii) Which of the following is an improper integral of 2nd kind.
0 3 1
1 1 πt
(I)  dx (II)  2 dx (III)  tan dt
−1
x 2
x −1 0
2
A. I and III only B. III only
C. I only D. II and III only
1 sin θ
(iii) The integral  dθ is ………………..
0 θ
A. improper integral of 1st kind. B. improper integral of 2nd kind.
C. improper integral of mixed kind. D. none of these.
∞
1
(iv) The integral  dx is ………………..
0
x
A. improper integral of 1st kind. B. improper integral of 2nd kind.
C. improper integral of mixed kind. D. none of these.
 Ch 01: Improper Integrals of 1st and 2nd Kinds 3

IMPROPER INTEGRAL OF THE FIRST KIND

 Definition
Assume that f ∈ R[ a, b ] for every b ≥ a . Define a function I on [a, ∞) as
follows:
b
I (b) =  f ( x ) dx for b ≥ a .
a
∞
The integral  a
f ( x ) dx is said to converge if the lim I (b) exists (finite).
b→∞
∞
Otherwise,  a
f dx is said to diverge.
If the lim I (b) exists and equals A , the number A is called the value of the
b→∞
∞
integral and we write  a
f dx = A .

 Remark
∞ ∞
 If  f dx is convergent(divergent), then  f dx is convergent(divergent)
a c
for c > a.
∞ ∞
 If  f dx is convergent (divergent), then  f dx is convergent (divergent)
c a
for a < c if f in bounded in [ a , c ] .

 Example
∞
1
Consider and integral x 1
2
dx . Discuss its convergence or divergence.

Solution
b
1
Let I (b) =  dx , where b ≥ 1 .
1
x2
b b
−1 b 1 1
Then I (b) =  x dx = − x −2
=− =1− .
1
1 x1 b
 1
Now lim I (b) = lim 1 −  = 1.
b→∞ b→∞
 b
∞ 1
Hence  2 dx is convergent.
1 x

 Example
∞
Consider and integral  1
x − p dx , where p is any real number. Discuss its
convergence or divergence.
4 Ch 01: Improper Integrals of 1st and 2nd Kinds

Solution
b
Let I (b) =  x − p dx , where b ≥ 1 .
1
b b
x1− p 1 − b1− p
Then I (b ) =  x dx =
−p
= if p ≠ 1.
1
1 − p 1
p −1
If b → ∞ , then
∞ if 1 > p,
b1− p → 
 0 if 1 < p.
Thus, we have
 ∞ if p < 1,
1 − b1− p 
lim I (b) = lim = 1
 p − 1 if p > 1.
b →∞ b →∞ p − 1

b
Now if p = 1 , we get  1
x −1 dx = log b → ∞ as b → ∞ .
diverges if p ≤ 1,
∞

Hence, we concluded:  x − p dx =  1
1  p −1 if p > 1.

 Review
If lim f ( x) = L , then for every sequence {xn } such that xn → a when n → ∞ ,
x →a
one has lim f ( xn ) = L .
n→∞
 Example
∞
Is the integral  sin 2π x dx converges or diverges?
0
Solution:
b
Consider I (b) =  sin 2π x dx , where b ≥ 0 .
0
b b
− cos 2π x 1 − cos 2π b
We have  sin 2π x dx = = .
0
2π 0 2π
1
Consider bn = n and cn = n + . Clearly bn → ∞ and cn → ∞ as n → ∞ .
2
But cos 2π bn → 1 and cos 2π cn → −1 as n → ∞ .
Thus limcos 2π b doesn’t exist and hence given integral is divergent.
b→∞
 Ch 01: Improper Integrals of 1st and 2nd Kinds 5

 Exercises
∞ 1
 Show that  a xp
dx converges if p > 1 .
0 0
 Evaluate: (i)  −∞
sin x dx (ii)  −∞
e x dx
 Note
a ∞
If 
−∞
f dx and a
f dx are both convergent for some value of a , we say
∞
that the integral  −∞
f dx is convergent and its value is defined to be the sum
∞ a ∞


−∞
f dx = 
−∞
f dx +  f dx .
a

The choice of the point a is clearly immaterial.

∞ b
If the integral 
−∞
fdx converges, its value is equal to the limit: lim 
b →+ ∞ − b
f dx .
For improper integral of first kind we will discuss the results for integral of
∞
the type a
f dx . The results for other cases can be derived in a similar manner.

 Exercises
∞
Evaluate the improper integral −∞
e x dx .
 MCQ
dx ∞
(i) For what value of m the integral
x m +1
is convergent.  1

A. m > 1 B. m ≤ 1 C. m > 0 D. m ≥ 0
(ii) Which of the following integrals is divergent.
3 3
∞ dx ∞ dt ∞ − ∞
A.  2 B.  α +1 , α > 0 C.  z 2 dz D.  x 2 dx
2 x 1 t 1 1
∞
(iii) If  2
fdx is convergent then ……………… is convergent.
∞ ∞ ∞ ∞
A.  0
fdx B.  1
fdx C. 3
fdx D. 
−2
fdx

 Review:
 A function f is said to be increasing, if for all x1 , x2 ∈ D f (domain of f )
and x1 ≤ x2 implies f ( x1 ) ≤ f ( x2 ) .
 A function f is said to be bounded if there exist some positive number
µ such that f (t ) ≤ µ for all t ∈ D f .
 If f is define on [a, +∞) and lim f ( x) exists then f is bounded on
x→∞
[a, +∞) .
6 Ch 01: Improper Integrals of 1st and 2nd Kinds

b c b
 If f ∈ R[ a, b ] and c ∈ [ a, b ] , then 
a
f dx =  f dx +  f dx .
a c
b
 If f ∈ R[ a, b ] and f ( x) ≥ 0 for all x ∈ [ a, b ] , then 
a
f dx ≥ 0 .
 If f is monotonically increasing and bounded on [a, +∞) , then
lim f ( x) = sup f ( x ) .
x →∞ x∈[ a ,∞ )
b b
 If f , g ∈R[a, b] and f ( x) ≤ g ( x) for all x ∈ [ a , b ] , then 
a
f dx ≤  g dx .
a

 Theorem
Suppose that f ∈R[a, b] for every b ≥ a . Assume that f ( x) ≥ 0 for each
∞
x ≥ a . Then  a
f ( x ) dx converges if, and only if, there exists a constant M > 0
such that
b

 f ( x) dx
a
≤ M for every b ≥ a .

Proof
b
Let I (b) =  f dx for b ≥ a .
a
∞
First suppose that  a
f ( x ) dx is convergent, then lim I (b) exists, that is, I (b)
b→+∞

is bounded on [a, +∞) .

So there exists a constant M > 0 such that
I (b) < M for every b ≥ a.
b
As f ( x) ≥ 0 for each x ≥ a , therefore  f ( x) dx ≥ 0 .
a
b
This gives I (b) =  f ( x ) dx ≤ M for every b ≥ a .
a
Conversely, suppose that there exists a constant M > 0 such that
b

 f ( x)dx
a
≤ M for every b ≥ a .

This give I (b) ≤ M for every b ≥ a , that is, I is bounded on [a, +∞) .
Now for b2 ≥ b1 > a , we have
b2 b1 b2

I (b2 ) =  f ( x) dx =  f ( x)dx +  f ( x) dx
a a b1
 Ch 01: Improper Integrals of 1st and 2nd Kinds 7

b1 b2

≥  f ( x ) dx = I (b1 ) ,
a
∵  f ( x) dx ≥ 0 as
b1
f ( x) ≥ 0 for all x ≥ a

.
This gives I is monotonically increasing on [a, +∞) .
As I is monotonically increasing and bounded on [a, +∞) , therefore lim I (b)
b→∞
∞
exists, that is,  a
f dα converges.

 Theorem: (Comparison Test)

Assume f ∈R[a, b] for every b ≥ a . If 0 ≤ f ( x) ≤ g ( x) for every x ≥ a and
∞ ∞ ∞ ∞
a
g dx converges, then  a
f dx converges and we have a
f dx ≤ 
a
g dx .
Proof
b b
Let I1 (b) =  f dx and I 2 (b) =  g dx , b ≥ a.
a a
Since 0 ≤ f ( x) ≤ g ( x) for every x ≥ a , therefore
b b

 f dx
a
≤  g dx ,
a

that is, I1 (b) ≤ I 2 (b) , b ≥ a …………………. (i)

∞
Since  g dx
a
converges, there exists a constant M > 0 such that
b

 g dx ≤ M ,
a
b ≥ a.

That is, I 2 (b) ≤ M , b ≥ a …………………(ii)

From (i) and (ii), we have I1(b) ≤ M for every b ≥ a .
∞
This implies  a
f dx converges, that is, lim I1 (b) exists and is finite.
b→∞
So we have lim I1 (b) ≤ lim I 2 (b) ≤ M ,
b→∞ b→∞
∞ ∞
this gives  f dx ≤  g dx .
a a
 Remark
∞ ∞
In comparison test, if 
a
f dx is divergent, then a
g dx is divergent.
 Example
∞
sin 2 x
Is the improper integral  2 dx convergent or divergent?
1
x
8 Ch 01: Improper Integrals of 1st and 2nd Kinds

Solution:
Since sin2 x ≤ 1 for all x ∈ [1, +∞ ) ,
sin 2 x 1
therefore ≤ 2 for all x ∈ [1, +∞ ) .
x2 x
∞ 2 ∞
sin x 1
This gives  2 dx ≤  2 dx .
1
x 1
x
∞ ∞
1 sin 2 x
Now  2 dx is convergent, therefore  2 dx is convergent.
1
x 1
x

 MCQs
(i) A function f is said to be bounded if there exist a positive number α such that
………….. for all t ∈ Df (domain of f )
A. f (t ) ≤ α B. f (t ) ≤ α C. f (t ) > α
D. f (t ) > α
(ii) If f : [ a , b ] → (0, ∞ ) is a bounded function then
b ∞
A. 
a
f (t )dt ≥ 0 B. 
a
f (t )dt ≥ 0 C. f (t ) ≥ µ for µ ∈ ℝ D. None of these

 Review:
 For all a , b , c ∈ ℝ , a − b < c ⇔ b − c < a < b + c or a − c < b < a + c .
 If lim f ( x) = m , then for all real ε > 0 , there exists N > 0 such that
x→∞

f ( x) − m < ε whenever x > N .

∞ ∞
 If  a
f dx converges(diverges), then 
N
f dx converges(diverges) if
N > a.
∞ ∞
 If  N
f dx is convergent (divergent), then a
f dx is convergent
(divergent) for a < N if f is bounded in [ a , N ] .

 Theorem (Limit Comparison Test)

Suppose that f , g ∈R[a, b] for every b ≥ a , where f ( x) ≥ 0 and g ( x) ≥ 0
for x ≥ a . If
f ( x)
lim = 1,
x →∞ g ( x )

∞ ∞
then  f dx
a
and  g dx both converge, or both diverge.
a

Proof
f ( x)
Suppose lim = 1 , then for all real ε > 0 , we can find some N > 0 , such
x →∞ g ( x )

that
 Ch 01: Improper Integrals of 1st and 2nd Kinds 9

f ( x)
−1 < ε ∀ x > N ≥ a.
g ( x)
f ( x)
 1− ε < <1+ ε ∀ x > N ≥ a.
g ( x)
1
If we choose ε = , then we have
2
1 f ( x) 3
< < ∀ x > N ≥ a.
2 g ( x) 2
This implies g ( x) < 2 f ( x) …..…..(i) and 2 f ( x) < 3g ( x) ……....(ii)
∞ ∞
From (i)  g dx < 2  f dx ,
N N
∞ ∞
so if 
a
f dx converges, then  f dx converges and hence by comparison test
N
∞ ∞
we get  g dx is convergent, which implies  g dx is convergent.
N a
∞ ∞
Now if  g dx diverges, then  g dx diverges and hence by comparison test
a N
∞ ∞
we get 
N
f dx is divergent, which implies  f dx is divergent.
a
∞ ∞
From (ii), we have 2  f dx < 3 g dx ,
N N
∞ ∞
so if  g dx converges, then  g dx converges and hence by comparison test
a N
∞ ∞
we get 
N
f dx is convergent, which implies  f dx is convergent.
a
∞ ∞
Now if  f dx diverges, then  f dx diverges and hence by comparison test
a N
∞ ∞
we get  g dx is divergent, which implies  g dx is divergent.
N a
∞ ∞
 The integrals  f dx
a
and  g dx
a
converge or diverge together.
10 Ch 01: Improper Integrals of 1st and 2nd Kinds

 Note
f ( x)
The above theorem also holds if lim = c , provided that c > 0 . If c = 0 ,
x →∞ g ( x )
∞
we can only conclude that convergence of  a
gdx implies convergence of
∞
a
fdx .

 Questions
(i) Suppose f ( x) and g ( x) are positive integrable functions for x > a . If
∞ ∞
f ( x)
lim = c , where c > 0 , then  f ( x ) dx and  g ( x ) dx both
x →∞ g ( x )
a a
converge or both diverge.
(ii) Suppose f ( x) and g ( x) are positive integrable functions for x > a . If
f ( x) ∞
lim = 0 , then convergence of  g ( x) dx implies convergence of
x →∞ g ( x ) a

∞
a
f ( x ) dx .
(iii) Suppose f ( x) and g ( x) are positive integrable functions for x > a .. If
f ( x) ∞
lim = ∞ , then convergence of  f ( x )dx implies convergence of
x →∞ g ( x ) a

∞
a
g ( x ) dx .

 Example
∞

e
−x
Prove that, for every real p , the integral x p dx converges.
1
Solution:
1
Let f ( x) = e− x x p and g ( x) = .
x2
f ( x) e− x x p
Now lim = lim
x →∞ g ( x ) x →∞ 1
x2
f ( x) x p +2
 lim − x p+2
= lim e x = lim x = 0 . (find this limit yourself)
x →∞ g ( x ) x →∞ x→∞ e

∞ ∞
1
Since  2 dx is convergent, therefore the given integral e
−x
x p dx is also
1
x 1
convergent.
 Ch 01: Improper Integrals of 1st and 2nd Kinds 11

 Remark
∞ ∞
It is easy to show that if  f dx and  g dx are convergent, then
a a
∞
•  a
( f ± g ) dx is convergent.
∞
•  a
cf dx , where c is some constant, is convergent.

 Review
 If lim f ( x) = m , then for all real ε > 0 , there exists real N > 0 such that
x→∞

f ( x ) − m < ε whenever x > N .

 A sequence {an } is said to be convergent if there exist a number l such
that for all ε > 0 , there exists a positive integer n0 (depending on ε ) such
that
an − l < ε whenever n > n0 .
The number l is called limit of the sequence and we write lim an = l .
n→∞

 A sequence {an } is said to be Cauchy if for all ε > 0 , there exists a

positive integer n0 such that
an − am < ε whenever n, m > n0 .
 A sequence of real numbers is Cauchy if and only if it is convergent.
 Theorem (Cauchy condition for infinite integrals)
∞
Assume that f ∈R[a, b] for every b ≥ a . Then the integral  f dx converges
a
if, and only if, for every ε > 0 there exists a B > 0 such that c > b > B implies
c

f
b
dx < ε .

Proof
∞ b
Let  f dx be convergent, that is,
a
lim  f dx exists and assume it to be A.
b →∞
a
Then for all ε > 0 , there exists B > a such that
b
ε
 f dx − A
a
<
2
for every b > B . ………..(i)

Also for c > b > B ,

c
ε
 f dx − A
a
<
2
…………….. (ii)
12 Ch 01: Improper Integrals of 1st and 2nd Kinds

c b c
As we know  f dx =  f dx +  f dx , this gives
a a b
c c b c b

 f dx
b
=  f dx −  f dx
a a
=  f dx − A + A −  f dx
a a
c b
ε ε
≤  f dx − A
a
+ A −  f dx <
a
2
+
2
=ε ,

c
  f dx
b
< ε when c > b > B .

Conversely, assume that the Cauchy condition holds.

For a positive integer n ≥ a , define
n
an =  f dx .
a
Consider n, m such that n > m > B , then
n m m n m
an − am =  f dx −  f dx
a a
=  f dx +  f dx −  f dx
a m a
n
=  f dx
m
<ε (by Cauchy condition).

This gives us that the sequence {an } is a Cauchy sequence.

This implies {an } is convergent and consider that lim an = A .
n→∞
Then for given ε > 0 , choose B so that
ε
an − A <whenever n ≥ B .
2
Also, for ε > 0 , we can have (by Cauchy condition)
c
ε
 f dx
b
<
2
if c > b > B .

Now if n, a, b > B such that b ≥ B + 1 , then we have

b n b

 f dx − A
a
=  f dx − A +  f dx
a n
b
ε ε
≤ an − A +  f dx
n
<
2
+
2
=ε .

b ∞
This gives us lim  f dx = A , that is,  f dx is convergent.
b →∞
a a
This completes the proof.
 Ch 01: Improper Integrals of 1st and 2nd Kinds 13

 Question:
Suppose f ∈R[a, b] for every b ≥ a and for every ε > 0 there exists a B > 0
c ∞
such that 
b
f dx < ε for b, c > B , then a
f dx is convergent.

 Example:
∞
sin x
Use Cauchy criterion to prove that 
1
x
dx is convergent.

Proof.
2 2
Let ε > 0 be an arbitrary and take B = such that c > b > B = . Then
ε ε
c c c
sin x − cos x cos x
b x dx = x b − b x 2 dx .
This gives us
c c
sin x cos b cos c cos x
b x dx = b − c − b x2 dx
c
cos b cos c cos x
≤
b
+
c
+ b x2 dx by triangular inequality

c b b
cos b cos c cos x
≤
b
+
c
+  2 dx
b
x
∵  f ( x)dx ≤ 
a a
f ( x) dx .
c
1 1 1
≤ + +  2 dx ∵ cos x ≤ 1 ∀ x ∈ ℝ .
b c bx
c
1 1 1 1 1 1 1
= + − = + + −
b c xb b c b c
2 ε 1 ε
= < 2⋅ =ε ∵ < .
b 2 b 2
2
That is, we have proved that for all ε > 0 , there exists B = such that
ε
c
sin x
b x dx < ε whenever c > b > B .
∞
sin x
Hence by Cauchy criterion 1 x dx is convergent.
14 Ch 01: Improper Integrals of 1st and 2nd Kinds

 Absolutely convergent
An improper integral
∞ ∞
 f dx is said to converge absolutely if 
a a
f dx

converges.

 Conditionally convergent
∞ ∞
An improper integral  f dx is said to be converge conditionally if  f dx
a a
∞
converges but a
f dx is divergent.

 Note: The definition of absolutely and conditionally convergence is the same

as above for other type of improper integrals.
 Theorem
∞ ∞
If f ∈R[a, b] for every b ≥ a and if 
a
f dx converges, then  f dx also
a
converges.
Proof
∞ ∞
Let 
a
fdx be absolutely convergent, i.e., a
f dx is convergent.
It is easy to see that
0 ≤ f ( x ) − f ( x ) ≤ 2 f ( x ) for all x ≥ a .
∞
Since we have given that 
a
f dx is convergent, therefore by comparison test,
∞
we have (
a
f − f ) dx converges.
∞ ∞
Now difference of 
a
f dx and (
a
f − f ) dx is convergent,
∞ ∞ ∞
that is, 
a
f dx −  ( f − f ) dx =  f dx is convergent.
a a

 Note: The converse of the above theorem doesn’t hold in general.

∞
sin x
For example: The integral  dx is convergent (prove yourself) but
1
x
∞
sin x

1
x
dx is divergent (it is hard to prove).
 Ch 01: Improper Integrals of 1st and 2nd Kinds 15

 Remarks:
 The above theorem can be stated as “an absolutely convergent integral is
convergent”.

 Questions
∞
 Use Cauchy criterion to prove that if an improper integral a
f ( x)dx is
absolutely convergent then it is convergent.
∞
 Show that 1
exp(− x 2 )dx is convergent.
∞ 1 + e− x
 Show that 1 x
dx is divergent.

 Review
 A function f ( x) is bounded for x ≥ a if there exist some positive number
K such that f ( x) ≤ K for all x ≥ a .
∞ ∞
 An integral a
f ( x)dx is said to be absolutely convergent if a
f ( x) dx
is convergent.
∞
 Let f ∈ R(a, b) for each b ≥ a . An integral  a
f dx converges if, and
b
only if, there exists a constant M > 0 such that a
fdx ≤ M for every
b ≥ a.

The following theorem provide useful test of convergence for product of

function.

 Theorem
If f ( x) is bounded for all x ≥ a , integrable on every closed subinterval of
∞
[a, ∞) (i.e. f ∈R[a, b] for each b ≥ a ) and a
g ( x)dx is absolutely convergent,
∞
then 
a
f ( x) g ( x)dx is absolutely convergent.
Proof
Since f ( x) is bounded for all x ≥ a , there exists K > 0 such that
f ( x ) ≤ K for all x ≥ a . …… (i)
∞ ∞
Since a
g ( x)dx is absolutely convergent, that is,  a
g ( x) dx is convergent,
there exists M > 0 such that
b
 a
g ( x) dx ≤ M for all b ≥ a . ……. (ii)
Now
16 Ch 01: Improper Integrals of 1st and 2nd Kinds

b b

 a
f ( x) g ( x ) dx =  f ( x ) g ( x) dx
a
b
≤ K  g ( x ) dx from (i)
a
≤ KM for all b ≥ a by using (ii).
b b
Hence 
a
f ( x) g ( x ) dx is convergent, this implies  f ( x) g ( x) dx is absolutely
a
convergent.

 Review
 Second Mean Value Theorem (Bonnet’s theorem): If f , g ∈ R[a, b] and
f is monotonic on [ a, b] , then there exist point c ∈ [ a, b] such that
b c b

  f ( x) g ( x)dx = f (a) g ( x)dx + f (b) g ( x)dx.

a a c
 A function f ( x) is bounded for x ≥ a if there exist some positive number
K such that f ( x) ≤ K for all x ≥ a .
b
 An integral 
a
f ( x ) dx converges if, and only if, for every ε > 0 there
c
exists a B > 0 such that b
f ( x) dx < ε when c > b > B.
 If lim f ( x) = m , then for all real ε > 0 , there exists real N > 0 such that
x→∞

 f ( x ) − m < ε whenever x > N .

 Theorem (Abel)
∞
If f ( x) is bounded and monotone for all x ≥ a and 
a
g ( x)dx is convergent,
∞
then 
a
f ( x) g ( x)dx is convergent.
Proof
As f is bounded and monotone on [a, ∞) , so it is integrable on [ a, b] , b > a .
Also g is integrable on [ a, b] for b > a .
By using second mean value theorem, we have
c c0 c

 f ( x) g ( x)dx = f (b)  g ( x)dx + f (c)  g ( x)dx, ….. (i)

b b c0

where a < b < c0 < c.

Since f is given to be bounded on [a, ∞) , there exists positive number K
such that
 Ch 01: Improper Integrals of 1st and 2nd Kinds 17

f ( x) ≤ K for x ≥ a .
In particular:
f (b) ≤ K and f (c) ≤ K . …….. (ii)
∞
Also a
g ( x)dx is convergent, by Cauchy criterion, for all ε > 0 , there exist
positive number B such that
c
ε
 g ( x)dx < 2 K
b
for b, c > B .

In particular:
c0 c
ε ε
 g ( x)dx < 2 K
b
and  g ( x)dx < 2K
c0
……… (iii)

From (i), (ii) and (iii), if b, c > B , we have

c c0 c

 f ( x) g ( x)dx
b
= f (b)  g ( x)dx + f (b)  g ( x)dx
b c0

c0 c
≤ f (b)  g ( x)dx +
b
f (b)  g ( x)dx
c0

ε ε
< K⋅ +K⋅ = ε.
2K 2K
∞
Hence by Cauchy criterion, we have that  a
f ( x) g ( x)dx is convergent.

 Theorem (Dirichlet)
If f ( x) is bounded, monotone for all x ≥ a and lim f ( x) = 0 . Also
x→∞
X ∞
a
g ( x)dx is is bounded for all X ≥ a , then  a
f ( x) g ( x)dx is convergent.
Proof
As f is bounded and monotone on [a, ∞) , so it is integrable on [ a, b] , b > a .
Also g is integrable on [ a, b] for b > a .
By using second mean value theorem, we have
c c0 c

 f ( x) g ( x)dx = f (b)  g ( x)dx + f (c)  g ( x)dx, ….. (i)

b b c0

where a < b < c0 < c.

X
Since  a
g ( x)dx is bounded for all X ≥ a , there exists positive number K
such that
18 Ch 01: Improper Integrals of 1st and 2nd Kinds

X

a
g ( x ) dx ≤ K for X ≥ a . …… (ii)
Now for a < b < c0 , we have
c0 b c0

a
g ( x)dx =  a
g ( x)dx +  g ( x)dx .
b
This gives
c0 c0 b
b
g ( x) dx =  a
g ( x ) dx −  g ( x) dx
a
c0 b
≤  a
g ( x ) dx +  a
g ( x) dx

≤ K + K by using (ii) as b, c0 > a .

c0
That is, 
b
g ( x ) dx ≤ 2 K . ………….. (iii)
c
Similarly,  c0
g ( x ) dx ≤ 2 A . …………….. (iv)
Also we have lim f ( x) = 0 , so for all ε > 0 , there exist B > 0 such that
x→∞
ε
f ( x) − 0 < for x > B .
4K
In particular, for b, c > B , we have
ε ε
f (b) < and f (c) < . …… (v)
4K 4K

From (i), (iii), (iv) and (v), if b, c > B , we have

c c0 c

 f ( x) g ( x)dx
b
= f (b)  g ( x)dx + f (b)  g ( x)dx
b c0

c0 c
≤ f (b)  g ( x)dx +
b
f (b)  g ( x)dx
c0

ε ε
< ⋅ 2K + ⋅ 2K = ε .
2K 2K
∞
Hence by Cauchy criterion, we have that a
f ( x) g ( x)dx is convergent.
 Ch 01: Improper Integrals of 1st and 2nd Kinds 19

 Example
∞
sin x
Prove that 
0
x
dx is convergent;

Solution:
sin x
Since → 1 as x → 0 , therefore 0 is not a point of infinite
x
discontinuity.
∞ 1 ∞ 1
sin x sin x sin x sin x
We write  dx =  dx +  dx and note that  dx is a proper
0
x 0
x 1
x 0
x
∞
sin x
integral. Therefore, it is enough to test the convergence of 
1
x
dx .

1
Denote f ( x) = and g ( x) = sin x , where x ≥ 1 .
x
1
Clearly f ( x) = ≤ 1 for x ≥ 1 implies f ( x) is bounded.
x
1 1
Now for x1 ≥ x2 ≥ 1 , we have ≤ , that is, f ( x1 ) ≤ f ( x2 ) . This gives us
x1 x2
f ( x) is decreasing for all for all x ≥ 1 .
1
Also lim f ( x) = lim = 0 .
x →∞ x →∞ x
X X
Now  g ( x) dx
1
=  sin x dx
1

= − cos X + cos(1) ≤ cos X + cos(1) < 2

X
This gives  g ( x) dx
1
is bounded for every X ≥ 1.
∞ ∞
sin x
Hence by Dirichlet theorem  f ( x) g ( x) dx = 
1 1
x
dx is convergent.

 Example
∞
Discuss the convergence of  sin x 2 dx .
1
1
Solution: We write sin x 2 = ⋅ 2 x ⋅ sin x 2 , i.e.
2x
∞ ∞
1
1 1 2 x ⋅ 2 x ⋅ sin x dx
2 2
sin x dx =
20 Ch 01: Improper Integrals of 1st and 2nd Kinds

1
Take f ( x) = and g ( x) = 2 x sin x 2 , where x ≥ 1.
2x
1
Note that f ( x) ≤ and f ( x) is decreasing for all for all x ≥ 1 , it gives f ( x)
2
1
is bounded and monotone for all x ≥ 1 . Also lim f ( x) = lim = 0 .
x →∞ x →∞ x
X X

 g ( x) dx  2 x sin x
2
Now = dx = − cos X 2 + cos(1) < 2 .
1 1
X
This gives 1 g ( x) dx is bounded for every X ≥ 1.
∞ ∞
1
Hence by Dirichlet’s theorem 1 f ( x ) g ( x) dx = 
1
2x
⋅ 2 x sin x 2 dx
∞

1 sin x
2
i.e. dx is convergent.

 Example
∞
sin x
0 e
−x
Discus the convergence of dx .
x
Solution:
sin x
Let f ( x) = e − x and g ( x) = , where x ≥ 0 .
x
As x ≥ 0 , then we have e x ≥ 1 (as exponential function is increasing).
1
This gives x ≤ 1 , that is, f ( x) ≤ 1 for all x ≥ 0 .
e
Also f ′( x) = −e − x < 0 for all x ≥ 0 . Hence we conclude that f ( x) is bounded
and monotonically decreasing for all x ≥ 0 .
∞ ∞
sin x
Also  g ( x ) dx =  dx is convergent (by previous example).
0 0
x
∞ ∞
sin x
Hence by Abel’s theorem 0 f ( x) g ( x )dx =  e − x
0
x
dx is convergent.

 Question
∞
sin x
Show that 0 (1 + x)α dx converges for α > 0 .
 Ch 01: Improper Integrals of 1st and 2nd Kinds 21

 Question
∞

0 e
−x
Show that cos x dx is absolutely convergent.

Solution
∞

0 e
−x −x −x
∵ e cos x < e and dx = 1

∴ the given integral is absolutely convergent. (comparison test).

 MCQs
∞
1. An improper integral a
f ( x)dx is conditionally convergent if it is
………………. but not ………………
(A) convergent; divergent (B) convergent; absolutely convergent
(C) divergent, convergent (D) divergent; absolutely convergent

2. If f is continuous and decreasing for x ≥ a and lim f ( x) = 0 , then the

x→∞
∞
integral a
f ( x)dx …………………..
(A) is convergent. (B) is divergent.
(C) may convergent. (D) absolutely convergent.

∞
f ( x)
3. The integral
1
x 2 
dx is ……………….. if f is bounded on [1, ∞) .

(A) convergent (B) divergent

(C) conditionally convergent (D) None of these
22 Ch 01: Improper Integrals of 1st and 2nd Kinds

IMPROPER INTEGRAL OF THE SECOND KIND

 Definition
Let f be defined on the half open interval ( a, b ] (having point of infinite
discontinuity at a ) and assume that f ∈R[ x, b] for every x ∈ ( a, b ] . Define a
function I on ( a, b ] as follows:
b
I ( x ) =  f dx if x ∈ ( a, b ]
x
b
If lim I ( x) exists then the integral
x→a +  f dx is said to be convergent. Otherwise,
a+
b

 f dx
a+
is said to be divergent.

If lim I ( x) = A , the number A is called the value of the integral and we write
x →a +
b

 f dx = A.
a+

Similarly, if f is defined on [a, b) (having point of infinite discontinuity at b)

x
and f ∈R[a, x] for all x ∈[a, b) then define I ( x ) =  f dx if x ∈[a, b) . If
a
b−

lim I ( x) exists (finite) then we say  f dx is convergent.

x→ b−
a
 Note
The tests developed to check the behaviour of the improper integrals of Ist
kind are applicable to improper integrals of IInd kind after making necessary
modifications.
 Example
b

x
−p
Discuss the convergence or divergence of dx for real p .
0
Solution:
f ( x) = x − p is defined on (0, b] and f , g ∈R[ x, b] for every x ∈ (0, b] .
b
I ( x ) =  u − p du if x ∈ (0, b]
x
b b b
u1− p b1− p − ε 1− p
u ε u
−p −p
du = lim du = lim = lim , ( p ≠ 1)
0+
ε →0
0+
ε →0 1 − p
ε
ε →0 1− p
 Ch 01: Improper Integrals of 1st and 2nd Kinds 23

 finite , p < 1
=
 infinite , p > 1
b
1
When p = 1 , we get  dx = log b − log ε → ∞ as ε → 0 .
ε x
b
 x
−1
dx also diverges.
0+

Hence the integral converges when p < 1 and diverges when p ≥ 1 .

 Note
c b−
If the two integrals  f dx and  f dx both converge, we write
a+ c
b− c b−

 f dx =  f dx +  f dx
a+ a+ c
The definition can be extended to cover the case of any finite number of sums.
We can also consider mixed combinations such as
b ∞ ∞

 f dx +  f dx
a+ b
which can be written as  f dx .
a+
 Question:
b b
dx dx
Prove that  ( x − a )n
a
and  ( b − x )n
a
converges if n < 1 . (see [4, page 490])

 Question
Examine the convergence of
1 1 1
dx dx dx
(i)  1 (ii) 0 x 2 (1 + x)2 (iii) x 1 1
0 x
3
1+ x 2
( ) 0
2
(1 − x ) 3

1
dx
Solution: (i) x 1
0
3
(1 + x ) 2

Here ‘0’ is the only point of infinite discontinuity of the integrand.

1 1
Let f ( x ) = 1 and take g ( x) = 1 .
x 3 (1 + x 2 ) x 3
f ( x) 1
Then lim = lim =1
x →0 g ( x ) x →0 1 + x 2
1 1
  0
f ( x) dx and 0
g ( x) dx have identical behaviours.
24 Ch 01: Improper Integrals of 1st and 2nd Kinds

1 1
dx dx
∵ x 1
converges ∴ x 1 also converges.
0
3
0
3
(1 + x ) 2

1
dx
(ii) 0 x (1 + x)2 2

Here ‘0’ is the only point of infinite discontinuity of the given integrand.
We have
1 1
f ( x) = 2 and take g ( x ) = .
x (1 + x) 2 x2
f ( x) 1
Then lim = lim 2
=1
x →0 g ( x ) x →0
(1 + x )
1 1
  0
f ( x) dx and 
0
g ( x) dx behave alike.
1
But n = 2 being greater than 1, the integral  0
g ( x) dx does not converge.
Hence the given integral also does not converge.
1
dx
(iii) x 1 1
0
2
(1 − x ) 3

Here ‘0’ and ‘1’ are the two points of infinite discontinuity of the integrand.
We have
1
f ( x) = 1 1
x 2 (1 − x ) 3
We take any number between 0 and 1, say 1 , and examine the convergence
2
1
2 1
of the improper integrals 
0
f ( x) dx and 
1
f ( x) dx .
2
1
2
1 1
To examine the convergence of x
0
1
2
(1 − x)
1
3
dx , we take g ( x) =
x
1
2

Then
f ( x) 1
lim = lim =1
x →0 g ( x) x→0 (1 − x ) 13
1 1
2 2
1 1
∵ x
0
1
2
dx converges ∴ 
0
1
x (1 − x) 2
1
3
dx is convergent.
 Ch 01: Improper Integrals of 1st and 2nd Kinds 25

1
1 1
To examine the convergence of 
1
1
x 2 (1 − x)
1
3
dx , we take g ( x ) =
(1 − x)
1
3
2
Then
f ( x) 1
lim = lim 1 = 1
x →1 g ( x) x→1 x 2
1 1
1 1
∵  (1 − x)
1
1
3
dx converges ∵ x
1
1
2
(1 − x)
1
3
dx is convergent.
2 2
1
Hence 0
f ( x) dx converges.

 Question
Show that the following improper integrals are convergent.
∞ ∞
2 1 sin 2 x
(i)  sin dx (ii)  2 dx
1
x 1
x
1 1
x log x
(iii)  2
dx (iv)  log x ⋅ log(1 + x) dx
0
(1 + x ) 0
1 1
Solution: (i) Let f ( x) = sin 2 and g ( x) = 2 .
x x
2
f ( x) sin 2 1x  sin y 
Then lim = lim 1 = lim   =1
x→∞ g ( x ) x →∞
2
y →0
 y 
x
∞ ∞
1
  f ( x) dx
1
and x
1
2
dx behave alike.
∞ ∞
1 1
∵  2 dx is convergent ∴  sin 2 dx is also convergent.
1
x 1
x
∞
sin 2 x
(ii)  2 dx
1
x
sin 2 x 1
Take f ( x ) = 2
and g ( x) = 2
x x
2
sin x 1
sin2 x ≤ 1  ≤ 2 ∀ x ∈ (1, ∞ )
x2 x
∞ ∞
1 sin 2 x
and  2 dx converges ∴  2 dx converges.
1
x 1
x
26 Ch 01: Improper Integrals of 1st and 2nd Kinds

 Note
1
sin 2 x sin 2 x
0 x 2 dx is a proper integral because lim
x →0 x2
= 1 so that ‘0’ is not a
∞
sin 2 x
point of infinite discontinuity. Therefore 0 x 2 dx is convergent.
1
x log x
(iii)  (1 + x)
0
2
dx

∵ log x < x , x∈ (0,1) ∴ x log x < x 2

x log x x2
 2
< 2
(1 + x ) (1 + x )
1 1
x2 x log x
Now  (1 + x )
0
2
dx is a proper integral, therefore  (1 + x )
0
2
dx is convergent.

1
(iv)  log x ⋅ log(1 + x) dx
0
∵ log x < x ∴ log( x + 1) < x + 1
 log x ⋅ log(1 + x) < x ( x + 1)
1
∵  x ( x + 1) dx
0
is a proper integral
1
∴  log x ⋅ log(1 + x) dx
0
is convergent.

 Note
a
1
(i)  x p dx
0
diverges when p ≥ 1 and converges when p < 1 .
∞
1
(ii) a x p dx converges iff p > 1.
 Questions
Examine the convergence of
∞ ∞ ∞
x 1 dx
(i) 
(1 + x ) 3
dx (ii) 1 (1 + x) x dx (iii) x 1 1
1 1
3
(1 + x ) 2

x 1
Solution: (i) Let f ( x) = 3
and take g ( x) = 2 .
(1 + x ) x
 Ch 01: Improper Integrals of 1st and 2nd Kinds 27

f ( x) x3
As lim = lim =1
x→∞ g ( x ) x→∞ (1 + x )3
∞ ∞
x 1
Therefore the two integrals  3
dx and  2 dx have identical
1
(1 + x ) 1
x
behaviour for convergence at ∞ .
∞ ∞
1 x
∵  2 dx is convergent ∴  3
dx is convergent.
1
x 1
(1 + x )

1 1 1
(ii) Let f ( x ) = and take g ( x) = = 3
(1 + x ) x x x x 2

f ( x) x
We have lim = lim =1
x →∞ g ( x ) x →∞ 1 + x

∞ ∞
1 1
and x
1
3
2
dx is convergent. Thus  (1 + x)
1 x
dx is convergent.

1
(iii) Let f ( x ) = 1 1
x 3
(1 + x ) 2

1 1
we take g( x) = 1 1
= 5
x 3 ⋅x 2
x 6
∞ ∞
f ( x) 1
We have lim
x →∞ g ( x )
= 1 and x
1
5
6
dx is divergent ∴  f ( x) dx is divergent.
1

 Question
∞
1
Show that
−∞
 1+ x 2
dx is convergent.

Solution: We have
∞
1 0 1 a
1 
 1 + x2 a→∞ −a 1 + x2 0 1 + x2 
dx = lim dx + dx
−∞  
a 1 a
1  a 1 
= lim   2
dx +  2
dx  = 2 lim   2
dx 
0 1+ x 1+ x  01+ x
a →∞ a →∞
0  
= 2 lim tan −1 x = 2   = π
a π
a →∞ 0
2
therefore the integral is convergent.
28 Ch 01: Improper Integrals of 1st and 2nd Kinds

 Question
∞
tan −1 x
Show that  2
dx is convergent.
0
1 + x
tan−1 x π tan x
−1

Solution: ∵ (1 + x ) ⋅ 2
= tan−1 x →
2
as x →∞ Here f ( x ) =
(1 + x ) 2 1+ x
2

∞
tan −1 x
∞
1 1
and g ( x) =
0 1 + x2 dx & 0 1 + x2 dx behave alike. 1+ x
2

∞
1
∵ 0 1 + x2 dx is convergent ∴ A given integral is convergent.

 Question
1
e− x
Show that 
0 1− x 4
dx is convergent.

2
Solution: ∵ e − x < 1 and 1+ x >1 for all x∈(0,1) .
e− x 1 1
∴ < <
1 − x4 (1 − x 2 ) (1 + x 2 ) 1 − x2
1 1−ε
1 1
Also 
0 1− x 2
dx = lim
ε →0 
0 1− x 2
dx

π
= lim sin −1 (1 − ε ) =
ε →0 2
1 −x
e
 
0 1− x 4
dx is convergent. (by comparison test)

References:
1. Tom M. Apostol, Mathematical Analysis, 2nd Edition, MA: Addison-Wesley, 1974.
2. S.C. Malik and Savita Arora, Mathematical Analysis, 2nd Edition, New Age
International, 1992.
3. S. Narayan and M.D. Raisinghania, Elements of Real Analysis, S. Chand &
Company, New Delhi, 2007.
4. D. Chatterjee, Real Analysis, 2nd Edition, PHI Learning Private Limited, Delhi, 2012.

For online resource related to the course

please visit: www.mathcity.org/atiq/sp19-mth322