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-III & IV
Course URL: www.mathcity.org/atiq/sp16-mth322

We discussed (in MTH321: Real Analysis I) Riemann-Stieltjes’s integrals of

b
the form  a
f d under the restrictions that both f and  are defined and bounded
b
on a finite interval [a, b] . The integral of the form  a
f d are called definite
integrals. To extend the concept, we shall relax some condition on definite integral
like f on finite interval or boundedness of f on finite interval.

 Definition
b
The integral  a
f d is called an improper integral of first kind if a    or
b    or both i.e. one or both integration limits is infinite.

 Definition
b
The integral  a
f d is called an improper integral of second kind if f ( x) is
unbounded at one or more points of a  x  b . Such points are called singularities of
f ( x) .

 Examples
 1 
1 1
  dx ,  dx and  ( x 2  1)dx are
0
1 x 2

x2 
examples of improper integrals of first kind.

1 1
1 1
  dx and  dx are examples of improper
1
x 0
2 x  1
integrals of second kind.

 Notations
We shall denote the set of all functions f such that
f R( ) on [a, b] by R( ; a, b) . When  ( x)  x , we
shall simply write R(a, b) for this set. The notation   on
[a, ) will mean that  is monotonically increasing on
[ a,  ) .
2 Ch 01: Improper Integrals of 1st and 2nd Kinds

IMPROPER INTEGRAL OF THE FIRST KIND

 Definition
Assume that f R( ; a, b) for every b  a . Keep a, and f fixed and define a
function I on [a,  ) as follows:
b
I (b)   f ( x) d ( x) if b  a ………… (i)
a
The function I so defined is called an infinite ( or an improper ) integral of first
 
kind and is denoted by the symbol  a
f ( x) d ( x) or by  a
f d .

The integral  a
f d is said to converge if the limit
lim I (b) ………… (ii)
b

exists (finite). Otherwise,  a
f d is said to diverge.
If the limit in (ii) exists and equals A , the number A is called the value of the

integral and we write  a
f d  A

 Example

Consider and integral  1
x  p dx , where p is any real number. Discuss its
convergence and divergence.
Solution
b
Let I (b)   x  p dx where b  1 .
1
b
x1 p 1  b1 p
b
Then I (b)   x dx 
p
 if p  1 .
1
1  p 1
p 1
If b   , then b1 p  0 for p  1 and b1 p   for p  1 . Therefore 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.

 Ch 01: Improper Integrals of 1st and 2nd Kinds 3

 Example

Is the integral  sin 2 x dx converges or diverges?
0

Solution:
b
Consider I (b)   sin 2 x dx , where b  0 .
0

 cos 2 x 1  cos 2 b
b b

We have  sin 2 x dx   .
0
2 0 2
Also cos 2 b  l as b   , where l has values between 1 and 1 , that is, limit is
not unique.

Therefore the integral  sin 2 x dx diverges.
0

 Note
a 
If 

f d and  f d
a
are both convergent for some value of a , we say that the

integral 

f d is convergent and its value is defined to be the sum
 a 



f d  

f d   f d
a

The choice of the point a is clearly immaterial.

 b
If the integral 

f d converges, its value is equal to the limit: lim
b   f d .
b

 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 
such that f (t )   for all t  D f .
 If lim f ( x) exists then f is bounded.
x

 If f R( ; a, b) and c  a, b , then

b c b
a
fd   fd   fd .
a c

 If f R( ; a, b) and f ( x)  0 for all x   a, b , then 

b
fd  0 .
a
 If f is monotonically increasing and bounded on [a,  ) , then
lim f ( x)  sup f ( x) .
x x[ a , )
4 Ch 01: Improper Integrals of 1st and 2nd Kinds

 Theorem
Assume that  is monotonically increasing on [a,  ) and suppose that

f R( ; a, b) for every b  a . Assume that f ( x)  0 for each x  a . Then  a
f d
converges if, and only if, there exists a constant M  0 such that
b

 f d
a
 M for every b  a .

Proof
b
Let I (b)   f d for b  a .
a

First suppose that  a
f d is convergent, then lim I (b) exists, that is, I (b) is
b
bounded.
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 d  0 .
a
b
This gives I (b)   f d  M for every b  a .
a
b
Conversely, suppose that there exists a constant M  0 such that  f d
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 d   f d   f d
a a b1
b1 b2

  f d  I (b1 )  f d  0 as f ( x)  0 for all x  a .

a b1

This gives I is monotonically increasing on [a,  ) .

As I is monotonically increasing and bounded on [a,  ) , therefore it is

convergent, that is  a
f d converges.
 Ch 01: Improper Integrals of 1st and 2nd Kinds 5

 Theorem: (Comparison Test)

Assume that  is monotonically increasing on [a,  ) and f R( ; a, b) for
 
every b  a . If 0  f ( x)  g ( x) for every x  a and  g d
a
converges, then  f d
a
converges and we have
 

 f d
a
  g d
a

Proof
b b
Let I1 (b)   f d and I 2 (b)   g d , b a.
a a
0  f ( x)  g ( x) for every x  a
 I1 (b)  I 2 (b) …………………. (i)


 g d
a
converges   a constant M  0 such that
b

 g d  M
a
, b  a …………………(ii)

From (i) and (ii) we have I1 (b)  M for every b  a .

 lim I1 (b) exists and is finite.
b

  f d
a
converges.

Also lim I1 (b)  lim I 2 (b)  M

b b
 
  f d   g d .
a a

 Example

sin 2 x
Is the improper integral  2 dx convergent or divergent?
1
x
Solution:
sin 2 x 1
Since sin x  1 for all x 1,   , therefore
2
 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
6 Ch 01: Improper Integrals of 1st and 2nd Kinds

 Theorem (Limit Comparison Test)

Assume that  is monotonically increasing on [a,  ) . Suppose that
f R( ; a, b) and that 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 d
a
and  g d
a
both converge or both diverge.

Proof
For all b  a , we can find some N  0 such that
f ( x)
1    x  N for every   0 .
g ( x)
f ( x)
 1   1 
g ( x)
1
Let   . Then we have
2
1 f ( x) 3
  .
2 g ( x) 2
 g ( x)  2 f ( x) …..…..(i) and 2 f ( x)  3g ( x) ……....(ii)
 
From (i)  g d
a
 2  f d ,
a
   
  g d
a
converges if  f d
a
converges and  g d
a
diverges if  f d
a
diverges.
 
From (ii) 2 f d  3 g d ,
a a
   
  f d
a
converges if  g d
a
converges and  g d
a
diverges if  f d
a
diverges.
 
 The integrals  f d
a
and  g d
a
converge or diverge together.

 Note
f ( x)
The above theorem also holds if lim  c , provided that c  0 . If c  0 , we
x  g ( x )

 
can only conclude that convergence of  g d
a
implies convergence of  f d .
a
 Ch 01: Improper Integrals of 1st and 2nd Kinds 7

 Example


e
x
For every real p , the integral x p dx converges.
1

1
This can be seen by comparison of this integral with 1 x2 dx .
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 x p2
 lim  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.
 Remark
 
It is easy to show that if  f d
a
and  g d are convergent, then
a

  ( f  g ) d
a
is convergent.

  cf d , where c is some constant, is convergent.
a

 Note
 
An improper integral  f d
a
is said to converge absolutely if 
a
f d converges.
 
It is said to be convergent conditionally if  f d
a
converges but 
a
f d diverges.

 Theorem

Assume   on [a,  ) . If f R( ; a, b) for every b  a and if 
a
f d

converges, then  f d
a
also converges.

Or: An absolutely convergent integral is convergent.

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

Proof
If x  a ,  f ( x)  f ( x)
 f ( x)  f ( x)  0  0  f ( x)  f ( x)  2 f ( x)

 
a
f  f  d converges.
 
Now difference of 
a
f d and 
a
f  f  d is convergent,

that is,  f d
a
is convergent.

 Remark
Every absolutely convergent integral is convergent.

 Theorem (Cauchy condition for infinite integrals)


Assume that f R( ; a, b) for every b  a . Then the integral  f d
a
converges if,

and only if, for every   0 there exists a B  0 such that c  b  B implies
c

f
b
d  

Proof
 b 
Let  f d
a
be convergent, that is lim  f d   f d .
b
a a B b c
Then  B  0 such that

b

 f d   f d
a a

2
for every b  B ………..(i)

Also for c  b  B ,

c


a
f d   f d 
a
2
…………….. (ii)
c b c
As we know  f d   f d   f d , this gives
a a b
c c b

 f d
b
  f d   f d
a a
c   b
  f d   f d   f d   f d
a a a a
 Ch 01: Improper Integrals of 1st and 2nd Kinds 9

 
c b
 
 
a
f d   f d 
a

a
f d   f d 
a
2

2


c
  f d
b
 when c  b  B .

Conversely, assume that the Cauchy condition holds.

an
Define an  a
f d if n  1,2,......

Consider n, m such that a  n, a  m  b  B , then

an am b an b am
an  am  a
f d  
a
f d   f d  
a b
f d   f d 
a
b
f d

an a m an am

 
b
f d  
b
f d   b
f d  
b
f d      2

This gives, the sequence an  is a Cauchy sequence  it is convergent.

Let lim an  A
n
c

Given   0 , choose B so that  f d
b

2
if c  b  B .

 a+N
and also that an  A  whenever a  n  B . a
2 B b c
Choose an integer N such that a  N  B .
Then, if b  a  N , we have
b a N b

 f d  A
a
 
a
f d  A  
a N
f d

b
 
 aN  A 
a N
 f d 
2

2



  f d  A
a
This completes the proof.
 Remarks

It follows from the above theorem that convergence of 
a
f d implies
b 
lim  f d  0 for every fixed   0 .
b b
However, this does not imply that f ( x)  0 as x   .

10 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 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 d if x   a, b ……….. (i)
x
The function I so defined is called an improper integral of the second kind and is
b b
denoted by the symbol  f (t ) d (t )
a
or  f d .
a
b
The integral  f d
a
is said to converge if the limit

lim I ( x) ……...(ii) exists (finite).

xa 
b
Otherwise,  f d
a
is said to diverge. If the limit in (ii) exists and equals A , the
b
number A is called the value of the integral and we write  f d  A .
a

Similarly, if f is defined on [a, b) and f R( ; a, x)  x [a, b) then

x
I ( x)   f d if x  [a, b) is also an improper integral of the second kind and is
a
b

denoted as  f d
a
and is convergent if lim I ( x) exists (finite).
x b

 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
f ( x)  x  p is defined on (0, b] and f R( x, b) for every x  (0, b] .
b
I ( x)   x  p dx if x  (0, b]
x
b b

x  x
p p
 dx  lim dx
 0
0 0
b
x1 p b1 p   1 p
 lim  lim , ( p  1)
 0 1  p  0 1 p

 Ch 01: Improper Integrals of 1st and 2nd Kinds 11

 finite , p  1

 infinite , p  1
b
1
When p  1 , we get  dx  log b  log    as   0 .
 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 d
a
and  f d
c
both converge, we write
b c b

a
 f d   f d   f d
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 d   f d
a b
which can be written as  f d .
a

 Assignment

Consider ( p)   e x x p1 dx , ( p  0) . Evaluate the convergence of this improper
0

integral.

 A Useful Comparison Integral

b
dx
  x  a n
a

We have, if n  1 ,
b b
dx 1
  x  a n
a 

(1  n)( x  a) n1 a 

1  1 1 
 
 (b  a ) n1  n1 
(1  n)  
1
Which tends to or   according as n  1 or n  1 , as   0 .
(1  n)(b  a)n1
Again, if n  1 ,
b
dx
 x  a  log(b  a)  log    
a
as   0 .
12 Ch 01: Improper Integrals of 1st and 2nd Kinds

b
dx
Hence the improper integral   x  a n
a
converges iff n  1 .

 Question
Examine the convergence of
1 1 1
dx dx dx
(i)  1 (ii) 0 x2 (1  x)2 (iii) x
  1  x  3
1 1
0 x
3
1 x 2
0
2

Solution
1
dx
(i) x
1  x 
1
3 2
0

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

We have
1
f ( x)  1
x 3 1  x2  
1
Take g ( x)  1
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.
1 1
dx dx
x converges  x also converges.
 
1 1
0
3
0
3
1  x2
1
dx
(ii)  x (1  x)
0
2 2

Here ‘0’ is the only point of infinite discontinuity of the given integrand.
We have
1
f ( x)  2
x (1  x) 2
1
Take g ( x)  2
x
f ( x) 1
Then lim  lim 1
1  x 
x0 g ( x) x0 2

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.
 Ch 01: Improper Integrals of 1st and 2nd Kinds 13

1
dx
(iii) x
1  x  3
1 1
2
0

Here ‘0’ and ‘1’ are the two points of infinite discontinuity of the integrand.
We have
1
f ( x)  1
x 2 1  x  3
1

We take any number between 0 and 1, say 1 , and examine the convergence of
2
1
2 1
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.

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
1
x 1
2 (1  x)
1
3
dx is convergent.
2 2
1
Hence  0
f ( x) dx converges.

 Question
1

 x 1  x 
m 1 n 1
Show that dx exists iff m , n are both positive.
0

Solution
The integral is proper if m  1 and n  1 .
The number ‘0’ is a point of infinite discontinuity if m  1 and the number ‘1’ is a
point of infinite discontinuity if n  1 .
Let m  1 and n  1 .
14 Ch 01: Improper Integrals of 1st and 2nd Kinds

We take any number, say 1 , between 0 & 1 and examine the convergence of the
2
1
2 1

 x 1  x   x 1  x 
m 1 n 1 m 1 n 1
improper integrals dx and dx at ‘0’ and ‘1’
0 1
2
respectively.
Convergence at 0:
We write
m 1 n 1 (1  x) n1 1
f ( x)  x (1  x)  and take g ( x) 
x1m x 1 m

f ( x)
Then  1 as x  0
g ( x)
1
2
1
As x
0
1 m
dx is convergent at 0 iff 1  m  1 i.e. m  0
1
2
x m1 1  x  dx is convergent at 0, iff m is +ive.
n 1
We deduce that the integral 
0
Convergence at 1:
m1 n1 x m1 1
We write f ( x)  x (1  x)  and take g ( x) 
(1  x)1n (1  x)1n
f ( x)
Then  1 as x  1
g ( x)
1
1
As  (1  x)
1
1 n
dx is convergent, iff 1  n  1 i.e. n  0 .
2
1

 x 1  x 
m 1 n 1
We deduce that the integral dx converges iff n  0 .
1
2
1

 x 1  x 
m 1 n 1
Thus dx exists for positive values of m , n only.
0
It is a function which depends upon m & n and is defined for all positive values of
m & n . It is called Beta function.
 Question
Show that the following improper integrals are convergent.
  1 1
2 1 sin 2 x x log x
(i)  sin dx (ii)  2 dx (iii)  dx (iv)  log x  log(1  x) dx
1
x 1
x 0
(1  x ) 2
0

Solution
1 1
(i) Let f ( x)  sin 2 and g ( x)  2
x x
 Ch 01: Improper Integrals of 1st and 2nd Kinds 15

2
f ( x) sin 2 1x  sin y 
then lim  lim 1  lim   1
x  g ( x ) x 
2
y 0
x y 
 
1
 
1
f ( x) dx and 1 x 2 dx behave alike.
 
1 2 1
1 x 2
dx is convergent  1 x dx is also convergent.
sin


sin 2 x
(ii) 1 x 2 dx
sin 2 x 1
Take f ( x)  2 and g ( x)  
x x
2
sin x 1
sin 2 x  1   2  x  1,  
x2 x
 
1 sin 2 x
and   dx converges   2 dx converges.
1
x 1
x

 Note
1
sin 2 x sin 2 x
0 x 2 dx is a proper integral because x0 x 2  1 so that ‘0’ is not a point of
lim

sin 2 x
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
 
1  x  1  x 
2 2

1
x2
Now  1  x 
0
2
dx is a proper integral.

1
x log x
  1  x 
0
2
dx is convergent.

1
(iv)  log x  log(1  x) dx
0
16 Ch 01: Improper Integrals of 1st and 2nd Kinds

log x  x  log( x  1)  x  1
 log x  log(1  x)  x ( x  1)
1 1

 x ( x  1) dx
0
is a proper integral   log x  log(1  x) dx is convergent.
0

 Note
a
1
(i) 0 x p dx diverges when p  1 and converges when p  1 .

1
(ii) x
a
p
dx converges iff p  1 .

 Questions
Examine the convergence of
  
x 1 dx
(i) 

dx (ii) 1 (1  x) x dx (iii) x
1  x 
3 1 1
1
(1 x ) 1
3 2

Solution
x 1
(i) Let f ( x)  and take g ( x )  .
(1  x)3 x2
f ( x) x3
As lim  lim 1
x g ( x) x (1  x)3
 
x 1
Therefore the two integrals  dx and 1 x2 dx have identical behaviour for
1
(1  x)3
convergence at  .
 
1 x
1 x2 dx is convergent  1 (1  x)3 dx is convergent.
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 (1  x) x dx is convergent.
1
(iii) Let f ( x) 
1  x 
1 1
3 2
x
1 1
we take g ( x)  1 1
 5
x 3 x 2
x 6
 Ch 01: Improper Integrals of 1st and 2nd Kinds 17

 
f ( x) 1
We have lim
x  g ( x )
 1 and x
1
5
6
dx is convergent   f ( x) dx is convergent.
1

 Question

1
Show that   1  x2 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   dx  0 1  x 2  a  0 1  x 2 
dx   2 lim  dx
0 1 x
a  2

a  
 2 lim tan 1 x  2    
a  0
2
therefore the integral is convergent.

 Question

tan 1 x
Show that  dx is convergent.
0
1  x 2

Solution
tan 1 x  1
(1  x )  2
 tan 1 x  as x   Here f ( x ) 
tan x
(1  x )
2
2 1 x
2

 
tan 1 x 1 1
0 1  x2 dx & 0 1  x 2 dx behave alike. and g ( x) 
1 x
2


1
1 x
0
2
dx is convergent  A given integral is convergent.

 Question


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

Solution


e
x x x
e cos x  e and dx  1
0
 the given integral is absolutely convergent. (Comparison test)
18 Ch 01: Improper Integrals of 1st and 2nd Kinds

 Question
e x
1
Show that 
0 1  x4
dx is convergent.

Solution
e x  1 and 1  x 2  1 for all x  (0,1) .
e x 1 1
  
1  x4 (1  x 2 )(1  x 2 ) 1  x2
1 1
1 1
Also 
0 1 x 2
dx  lim
 0 
0 1 x 2
dx


 lim sin 1(1   ) 
 0 2
1 x
e
 
0 1  x4
dx is convergent. (by comparison test)

References:
(1) Book
Mathematical Analysis
Tom M. Apostol (John Wiley & Sons, Inc.)