Improper Integrals

aft
Sharad Bhagwanrao Kadam

Assistant Professor

Department of Mathematics
Dr
Shri Shivaji College, Parbhani - 431 401.

[email protected]

April 24, 2021

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 1 / 67

Introduction

The concept of Riemann integrals requires that the range of integration is finite and the integrand

aft
remains bounded in that domain. In case the integrand f becomes infinite in the interval a 6 x 6 b,
i.e., f has points of infinite discontinuity (singular points) in [a, b], or the limits of integration a
Zb
or b (or both) become infinite, the symbol f dx is called an improper (or infinite or generalised)
a
integral. Thus
Z∞ Z∞ Z1 Z∞
dx dx dx dx
, , ,
1 x2 −∞ 1 + x2 0 x (1 − x) −1 x2
are examples of improper integrals.
Dr
The integrals which are not improper are called proper integrals. Thus
integral.
Z1

0
sin x
x
dx is a proper

It is assumed once for all that in a finite interval which encloses no point of infinite discontinuity
(singular point) the integrand is bounded and integrable.

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 2 / 67

Integration of Unbounded Functions with Finite Limits of Integration

Let a function f be defined in an interval [a, b] everywhere except possibly at a finite number of
points.
i) Convergence at the left-end : Let a be the only point of infinite discontinuity of f , the integral
Zb Zb

aft
f dx exists for every λ, 0 < λ < b − a. The improper integral f dx is defined as the limit
a+λ a+λ
Zb
of f dx when λ → 0+, so that
a+λ
Zb Zb
f dx = lim f dx (1)
a λ→0+ a+λ

Zb
Dr
If this limit exists and is finite, the improper integral
otherwise it is called divergent.
a
f dx is said to exist or converge (at a ),

ii) Convergence at the right-end : Let b be the only point of infinite discontinuity of f , the improper
integral is then defined by the relation
Zb Z b−µ
f dx = lim f dx, 0<µ<b−a (2)
a µ→0+ a

If the limit exists, the improper integral is said to be convergent (at b), otherwise it is called
divergent.

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 3 / 67

iii) Convergence at both the end-points : If the end-points a and b are the only points of infinite
discontinuity of f , then for any point c within the interval [a, b], the improper integral
Zb Zc Zb
f dx = f dx + f dx (3)
a a c

aft
If both the integrals in (3) exist in accordance with the definitions given above, the improper
integral converges, otherwise it is divergent.
The improper integral is also defined as
Zb Z b−µ
f dx = lim f dx (4)
a λ→0+ a+λ
µ→0+

The improper integral exists if the limit exists.

discontinuity of f , we put
Dr
iv) Convergence at interior points : If an interior point c, a < c < b, is the only point of infinite
Zb Zc

a
Zb
f dx = f dx + f dx
a c
Zb
The improper integral f dx is convergent if both the integrals on the R.H.S. exist in accordance
a
with the definitions given above.

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 4 / 67

Similarly if the function has a finite number of points of infinite discontinuity, c1 , c2 , . . . , cm within

aft
[a, b], where
a 6 c1 < c2 < . . . < cm 6 b,
Zb
the improper integral f dx is defined as
a
Zb Z c1 Z c2 Zb
f dx = f dx + f dx + . . . + f dx (5)
a a c1 cm

Dr
and is said to be convergent if all the integrals on the R.H.S. of equation (5) are convergent,
otherwise it is divergent.

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 5 / 67

Example

Z1 Z1
dx dx
Examine the convergence of 2
, √ .
0 x 0 1−x
Solution : 0 is the only point of infinite discontinuity of the integrand in [0, 1]. Thus,

aft
Z1 Z1
dx dx
= lim , 0<λ<1
0 x2 λ→0+ x2 λ
 
1
= lim −1 =∞
λ→0+ λ

Thus the improper integral is divergent.

Since 1 is the only point of infinite discontinuity of the integrand in [0, 1], we put

0
Dr
Z1
√
dx
1−x
= lim
µ→0+
h √
Z 1−µ
√
dx
1−x
0

= lim −2 1 − x
, 0<µ<1
i1−µ
µ→0+ 0
√
= − lim 2 ( µ − 1) = 2
µ→0+

Thus the improper integral exists and is equal to 2.

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 6 / 67

Example

Z2
dx
Examine the convergence of .

aft
0(2 x − x2 )
Solution : Both the end-points 0, 2 are the points of infinite discontinuity of the integrand and are
in fact the only such points in [0, 2]. Thus for any point, say 1, within [0, 2], we put
Z2 Z1 Z 2−µ
dx dx dx
= lim + lim
0 2 x − x2 λ→0+ λ x (2 − x) µ→0+ 1 x (2 − x)
 1  2−µ
1 x 1 x
= lim log + lim log
2 λ→0+ 2−x λ 2 µ→0+ 2−x 1
Dr=−

=∞
1
lim log
2 λ→0+
λ
2−λ
+
1
lim log
2 µ→0+
2−µ
µ

Thus, the given integral diverges.

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 7 / 67

 Zb
Comparison Tests for Convergence at a of f dx.
a

Zb
A necessary and sufficient condition for the convergence of the improper integral f dx at a,
a

aft
where f is positive in [a, b] is that there exists a positive number M, independent of λ, such that
Zb
f dx < M, 0 < λ < b − a.
a+λ
Zb
Proof : Given f is positive in [a, b]. We know that the improper integral f dx converges at a if for
Zb a

0 < λ < b − a, f dx tends to a finite limit as λ → 0 + .

a+λ Dr
Since f is positive in [a, b], the positive function of λ,
Zb
f dx is monotone increasing as λ decreases
a+λ
and will therefore tend to a finite limit if and only if it is bounded above, i.e., there exists a positive
number M independent of λ such that
Zb
f dx < M, 0 < λ < b − a.
a+λ

Hence, the proof.

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 8 / 67

Comparison Test I (Comparison of two integrals)

If f and g be two positive functions such that f (x) 6 g(x), for all x in [a, b], then
Zb Zb
(i) f dx converges, if g dx converges, and
a a

aft
Zb Zb
(ii) g dx diverges, if f dx diverges.
a a
Proof : Let f and g be both bounded and integrable in [a + λ, b], 0 < λ < b − a and a is the only
point of infinite discontinuity in [a, b]. Since f and g are positive and

f (x) 6 g(x), ∀ x ∈ [a, b]

Zb Zb

(i) Let
Zb
Dr ∴
a+λ
f dx 6
a+λ
g dx

g dx be convergent, so that there exists a positive number M such that

(6)

a
Zb
g dx < M, for 0 < λ < b − a
a+λ

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 9 / 67

 Thus, from inequality (6)

aft
Zb
f dx < M, for 0 < λ < b − a
a+λ
Zb
Hence f dx converges at a.
a
Zb Zb
(ii) Again, if f dx is divergent at a, then the positive function f dx is not bounded above
a a+λ
Zb Z b+λ
Dr
and therefore from equation (1),
divergent at a.
g dx is also not bounded above. Hence,
a+λ a
g dx is

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 10 / 67

Comparison Test II (Limit form)

f (x)
If f and g are two positive functions in [a, b] such that lim = l, where l is a non-zero finite
x→a+ g(x)
Zb Zb
number, then the two integrals f dx and g dx converge and diverge together at a.

aft
a a
Proof : Evidently l > 0.
f (x)
Let ε be a positive number such that l − ε > 0. Since lim = l, therefore there exists a
g(x) x→a+
neighbourhood (a, c), (a < c < b), such that for all x ∈ (a, c), we have

f (x)
−l <
g(x)
or

Now,
Dr (l − ) g(x) < f (x) < (l + ) g(x) (7)

(l − ) g(x) < f (x), ∀ x ∈ (a, c)

Zb Zc Zc Zb
So if f dx, i.e., if f dx converges at a then by comparison test I, (l − ) g dx and hence, g dx
a a a a
converges at a.

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 11 / 67

Again from inequality (7),
f (x) < (l + ) g(x), ∀ x ∈ (a, c)
Zb Zc Zc Zb

aft
So if f dx i.e., if
f dx diverges at a, then by comparison test, (I + ε) g dx and therefore, g dx
a a
Zb a
Zb a

diverges at a. It may similarly be shown that f dx converges and diverges with g dx.
a a
Hence, the two integrals behave alike.

Note: It can be easily shown that:

Z Z Z
f
(i) if

(ii) if
g

f
g
→ ∞ and
Zb

a
Dr
→ 0 and g dx converges, then f dx also converges, and

Z
g dx diverges, then f dx also diverges.

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 12 / 67

A Useful Comparison Integral

Zb
dx
The improper integral converges if and only if n < 1.
(x − a)n
a
Proof : It is a proper integral if n 6 0, and improper for other values of n, a is the only point of
infinite discontinuity of the integrand.

aft
Now for n , 1,
Zb Zb
dx dx
n
= lim , 0<λ<b−a
a (x − a) λ→0+ a+λ (x − a)n
 
1 1 1
= lim n−1
− n−1
λ→0+ 1 − n (b − a) λ


 1
, if n < 1
n−1
= (1 − n) (b − a)
Dr 
∞, if n > 1
Again, for n = 1
Zb Zb Zb
dx dx dx
= = lim
a (x − a)n a x−a λ→0+ a+λ x−a
= lim {log(b − a) − log λ} = ∞
λ→0+
Zb
dx
Thus, converges only for n < 1.
a (x − a)n

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 13 / 67

Example

Test the convergence of

Z1
dx

aft
√ .
0 1 − x3
Solution : Let
1 1 1
f (x) = √ =
1 − x3 (1 − x)1/2 (1 + x + x2 )1/2
1
Clearly, 1/2
is a bounded function and let M be its upper bound.
(1 + x + x2 )
M

Also
Z1
dx
(1 − x)1/2
0
is convergent.
Z1
Dr ∴ f (x) 6
(1 − x)1/2

dx
Therefore by comparison test, √ is convergent.
0 1 − x3

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 14 / 67

Example

Test the convergence of

Z π/2
sin x
dx.

aft
0 xp
Solution : For p 6 1, it is a proper integral. For p > 1, it is an improper integral, 0 being the point
of infinite discontinuity. Now
sin x 1 sin x
= p−1 ·
xp x x
sin x sin x
The function is bounded and 6 1.
x x

Also
Z π/2
dx
Dr ∴
sin x
xp

converges only if p − 1 < 1 or p < 2.

x
1
6 p−1

xp−1
0
Z π/2
sin x
Therefore by comparison test, dx converges for p < 2 and diverges for p > 2.
0 xp

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 15 / 67

Observations

i) If f is positive in a neighbourhood of a, then the integral

Zb
f dx
a

aft
converges at a if there exists a positive number n less than 1 and a fixed positive number M
such that
M
f (x) 6
(x − a)n
for all x in (a, b].
Zb
Also, f dx diverges if there exists a number n > 1 and a fixed positive number G such that
a

x in (a, b].
Dr f (x) >
G
(x − a)n

ii) If lim [(x − a)n f (x)] exists and is non-zero finite, then the integral
x→a+

Zb
f dx
a

converges if and only if n < 1.

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 16 / 67

Example

aft
Find the values of m and n for which the following integral converge:
Z1
e−m x xn dx
0

Solution : Let k be a number greater than 1 and e−m , for all m.

Z1 Z1
dx
In [0, 1], e−m xn 6 k xn , for all m, and xn dx = converges for −n < 1 or n > −1 only.
Z1 Dr 0 0 x
−n

Thus, e−m x xn dx converges only for n > −1, irrespective of the values of m.
0

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 17 / 67

Example

Find the values of m for which the following integral converge :

Z1
(log 1/x)m dx.

aft
0
Z1 Z 1/2 Z1
Solution : Putting (log 1/x)m dx = (log 1/x)m dx + (log 1/x)m dx,
0 0 1/2
0 and 1, respectively are the points of infinite discontinuity of the integrals on the right. Let

f (x) = (log 1/x)m .

Convergence at 0 :
Z 1/2

0
Dr
(log 1/x)m dx is a proper integral if m 6 0, for the integrand tends to a finite limit as x → 0 (1
if m = 0; 0 if m < 0). 0 is the only point of infinite discontinuity if m > 0.
For m > 0, take
1
g(x) = p , 0 < p < 1
x
so that
f (x)
= xp (log 1/x)m → 0 as x → 0
g(x)

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 18 / 67

 Z 1/2 Z 1/2
Also g dx converges, therefore (log 1/x)m dx converges.
Z0 1/2 0

Thus, (log 1/x)m dx converges for all m.

0
Convergence at 1 :
Z1

aft
(log 1/x)m dx is a proper integral if m > 0, and 1 is the only point of infinite discontinuity if
1/2
m < 0. For m < 0, let
1
g(x) =
(1 − x)−m
so that
log 1/x n
 
f (x)
= → 1 as x → 1.
g(x) 1−x

Hence, the two integrals

Z1
dx
−m
converges
Z1

1/2

if −m
Dr
f dx and

< 1
Z1

1/2

or
g dx behave alike.

m > −1 therefore
Z1
(log 1/x)m dx also converges if
1/2 (1 − x) 1/2
0>m> Z −1. 1
Hence, (log 1/x)m dx is convergent when 0 > m > −1.
0

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 19 / 67

Example

Z1
log x
Show that √ dx is convergent.
0 x √
Solution : Since log x/ x is negative in [0, 1], we take

aft
log x log 1/x
f (x) = − √ = √
x x

0 is the only point of infinite discontinuity.

Let
1
g(x) =
x3/4
so that

The integral
Dr
f (x)
g(x)
= x1/4 (log 1/x) → 0 as x → 0+

Z1
g dx =
Z1
dx
0 0 x3/4
is convergent at 0.
Z1 Z1
log 1/x log x
Therefore, √ dx and so √ dx is also convergent.
0 x 0 x

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 20 / 67

Example

Z2 √
x
Show that dx is divergent.
1 log x
Solution : Here 1 is the only point of infinite discontinuity.
Let

aft
√
x
f (x) =
log x
and
1
g(x) =
x−1
so that √ 3
f (x) (x − 1) x 2 x1/2 − 21 x−1/2
lim = lim = lim =1

The integral
x→1+ Dr
g(x) x→1+ log x

Z2

1
x→1+

dx
x−1
1/x

diverges and therefore

Z2 √
x
dx
1 log x
also diverges.

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 21 / 67

Example

Z π/2 
sinm x

Show that the integral dx exists if and only if n < m + 1.
0 xn
Solution : Let m
sinm x

aft

1 sin x
f (x) = = n−m ·
xn x x
Here as x → 0+, f (x) → 0 if n − m < 0, and f (x) → ∞ if n − m > 0.
Thus, it is a proper integral if n 6 m, and improper if n > m, 0 being the only point of infinite
discontinuity of f .
When n > m, let
1
g(x) = n−m
x
so that

Z π/2 Z π/2
dx
Dr
f (x)
g(x)
=

sin x m
x

→ 1 as x → 0+

Also g dx = converges if and only if n − m < 1 or n < m + 1, therefore

xn−m
Z π/2 0 0
sinm x
dx also converges if and only if n < m + 1 which includes the case n 6 m when the
0 xn
integral is proper.

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 22 / 67

Beta Function
Z1
Show that xm−1 (1 − x)n−1 dx exists if and only if m, n are both positive.
0
Solution : It is a proper integral for m > 1, n > 1. Here 0 and 1 are the only points of infinite
discontinuity; 0 when m < 1, and 1 when n < 1.

aft
1
Taking a number, say , between 0 and 1, we put
2
Z1 Z 1/2 Z1
xm−1 (1 − x)n−1 dx = xm−1 (1 − x)n−1 dx + xm−1 (1 − x)n−1 dx
0 0 1/2

and examine the two integrals on the right for convergence at 0 and 1 respectively.
Convergence at 0 : When m < 1.
Let
Dr
f (x) = xm−1 (1 − x)n−1 =

f (x)
g(x)
(1 − x)n−1
x1−m
and g(x) = 1−m
x

= (1 − x)n−1 → 1 as x → 0
1

Z 1/2 Z 1/2
dx
Also g dx = converges at 0 if and only if 1 − m < 1, i.e., m > 0.
x1−m
Z01/2 0

Thus xm−1 (1 − x)n−1 dx converges at zero if and only if m > 0.

0

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 23 / 67

Convergence at 1 : When n < 1.

xm−1
f (x) = xm−1 (1 − x)n−1 = .
(1 − x)1−n

aft
Let
1
g(x) =
(1 − x)1−n
then
f (x)
= xm−1 → 1 as x → 1
g(x)
Z1 Z1
dx
Also g dx = converges at 1, if and only if 1 − n < 1, or n > 0.
(1 − x)1−n

Thus

Hence
1/2
Z1

1/2
Z1
1/2
Dr
xm−1 (1 − x)n−1 dx converges at 1, if and only if n > 0.

xm−1 (1 − x)n−1 dx exists for positive values of m and n only.

0
Z1
This integral xm−1 (1 − x)n−1 dx, for m, n > 0 is called Beta function and is denoted by β(m, n).
0

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 24 / 67

Example
Z1
For what values of m and n is the integral xm−1 (1 − x)n−1 log x dx convergent?
0
Solution : The integrand is negative in [0, 1], therefore we shall test for convergence the integral
Z1 Z1
1

aft
−xm−1 (1 − x)n−1 log x dx, i.e. xm−1 (1 − x)n−1 log dx
0 0 x

1
Since 0 and 1 are the possible points of infinite discontinuity, therefore taking a number, say ,
2
between 0 and 1, we examine the integrals
Z 1/2 Z1
1 1
xm−1 (1 − x)n−1 log dx, xm−1 (1 − x)n−1 log dx
0 x 1/2 x

at 0 and 1 respectively.
Convergence at 0 :
Dr
It is a proper integral for m − 1 > 0, and improper for m 6 1, 0 being the only point of infinite
discontinuity, then for m 6 1, let

1 (1 − x)n−1 log 1/x 1

f (x) = xm−1 (1 − x)n−1 log = and g(x) = p
x x1−m x
Z 1/2
g dx is convergent if and only if p < 1. Also
0

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 25 / 67

 f (x) 1
= xp+m−1 (1 − x)n−1 log → 0
g(x) x
as x → 0+ if p + m − 1 > 0 or m > 1 − p > 0.
Z 1/2
So by comparison test, f dx converges if and only if m > 0.
0

aft
Convergence at 1 :
log 1/x (1 − x)n
Since lim = lim exists finitely when n > 0, therefore the integral is proper
x→1− (1 − x)1−n x→1 (1 − n)x
for n > 0 and improper for n < 0, 1 being the only singular point. For n < 0, let

1 xm−1 log 1/x 1

f (x) = xm−1 (1 − x)n−1 log = and g(x) =
x (1 − x)1−n (1 − x)q
Z1

1/2
Dr
g dx is convergent if and only if q < 1. Also

f (x)
g(x)
=
xm−1 log 1/x
(1 − x)1−n−q

tends to a finite limit as x → 1, if 1 − n − q 6 1, i.e., if n > −q > −1.

Z1
Thus, f dx converges at 1 if n > −1.
1/2
Hence, the given integral is convergent when m > 0, n > −1.

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 26 / 67

Example

Z π/2
Show that the integral log sin x dx is convergent and hence evaluate it.
0
Solution : Let f (x) = log sin x. As f is non-positive in [0, π/2], we consider the function (−f ) for

aft
testing convergence of the integral. 0 is the only point of infinite discontinuity of f . Let

1
g(x) = , m < 1,
xm
so that
f (x)
− = −xm log sin x → 0 as x→0
g(x)
Z π/2
dx
Also, we know that

Therefore,
Z π/2

0
0 xm
Dr
is convergent for m < 1.

− log sin x dx is convergent and so

Z π/2
Z π/2

0
log sin x dx is convergent.

To evaluate the integral, let I = log sin x dx. We have

0

sin 2x = 2 sin x cos x ⇒ log sin 2x = log 2 + log sin x + log cos x

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 27 / 67

 Z π/2 Z π/2 Z π/2 Z π/2
log sin 2x dx = log 2 dx + log sin x dx + log cos x dx
0 0 0 0
Z π/2

aft
π
= log 2 + I + log cos x dx
2 0
π
Putting 2 x = t in the first and x = − y in the last integral, we get
2
π
I= log 2 + I + I
2
π

Hence
Dr Z π/2

0
∴ I=−

log sin x dx = −
2
log 2

π
2
log 2

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 28 / 67

General Test for Convergence (Integrand may change sign)
Zb
Cauchy’s Test : The improper integral f dx converges at a if and only if to every ε > 0 there
a
corresponds δ > 0 such that
Z a+λ2

aft
f dx < ε, 0 < λ1 , λ2 < δ
a+λ1

Z a+λ2
[Notice that f dx tends to 0 as λ1 , λ2 → 0. ]
a+λ1
Zb Zb
Proof : The improper integral f dx is said to exist when lim f dx exists finitely.
a λ→0+ a+λ
Zb
Let F(λ) = f dx so that F(λ) is a function of λ.
a+λ Dr
According to Cauchy’s criterion for finite limits F(λ) tends to a finite limit as λ → 0 if and only if
for every  > 0 there corresponds δ > 0 such that for all positive λ1 , λ2 < δ.

|F (λ1 ) − F (λ2 )| < 

Zb Zb
⇒ f dx − f dx < ε
a+λ1 a+λ2
Z a+λ2
⇒ f dx < ε
a+λ1

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 29 / 67

Absolute Convergence
Zb Zb
The improper integral f dx is said to be absolutely convergent if |f | dx is convergent.
a a
Theorem : Every absolutely convergent integral is convergent, or
Zb Zb
f dx exists if |f | dx exists.

aft
a a
Zb
Proof : Since |f | dx exists, therefore by Cauchy’s test, for ε > 0 there corresponds δ > 0 such
a
that
Z a+λ2
|f | dx < ε, 0 < λ1 , λ2 < δ (8)
a+λ1

Also, we know that Dr Z a+λ2

a+λ1
f dx 6
Z a+λ2

a+λ1
|f | dx (9)

Hence, from equations (8) and (9)

Z a+λ2
f dx < ε, 0 < λ1 , λ2 < δ
a+λ1
Zb
Hence, f dx exists.
a
Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 30 / 67
Example

Z1
sin 1/x
Show that dx, p > 0, converges absolutely for p < 1.

aft
0 xp
Solution : Let
sin 1/x
f (x) = , p>0
xp
0 is the only point of infinite discontinuity, and f does not keep the same sign in any neighbour-
hood of 0. In [0, 1],
sin 1/x | sin 1/x| 1
|f (x)| = = < p
xp xp x

Also
Z1
dx
0 x
Dr
p converges if and only if p < 1.

Hence by comparison test, the integral

Z1

0
sin 1/x
xp
dx converges and so
Z1

0
sin 1/x
xp
dx converges
absolutely if and only if p < 1.

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 31 / 67

Infinite Range of Integration

We shall now consider the convergence of improper integrals of bounded integrable function with
infinite range of integration (a or b or both infinite.)
i) Convergence at ∞ : The symbol

aft
Z∞
f dx, x > a (10)
a
ZX
is defined as the limit of f dx, when X → ∞, so that
a
Z∞ ZX
f dx = lim f dx
Dr a X→∞ a

If the limit exists and is finite, then the improper integral (10) is said to be convergent, otherwise
it is said to be divergent.
Example : Examine for convergence
Z∞ Z∞ Z∞
x 1
2
dx, √ dx, sin x dx
0 1+x 1 x a

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 32 / 67

ii) Convergence at −∞ :
Zb
f dx, x6b (11)
−∞

aft
is defined by the equation
Zb Zb
f dx = lim f dx
−∞ X→−∞ X

If the limit exists and is finite then the integral (11) converges (or exists). Otherwise it diverges (or
does not exist).
iii) Convergence at both ends :
Z∞

is understood to mean
Dr Zc
−∞
f dx, ∀ x

Z∞
(12)

f dx + f dx (13)
−∞ c

where c is any real number.

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 33 / 67

If both the inegrals in (13) exist in accordance with the definition given above, then the integral
equation (12) converges, otherwise it is divergent.
Integral (12) can also be defined by the relation
Z∞ ZX

aft
f dx = lim f dx (14)
−∞ X→∞ Y
Y→−∞

iv) Integrals of unbounded functions with infinite limits of integration :

When the infinte range of integration includes a finite number of points of infinte discontinuity
Z of
∞
f , consider an interval [a, b] which contain all the points of discontinuity. The integral f dx is
−∞
then understood to mean

In
Dr Za

−∞
f dx +
Zb

a
f dx +
Z∞

b
f dx

Z ∞case all the integrals in (15) exist in accordance with the definitions given above, the integral
(15)

f dx converges, otherwise it is divergent.

−∞

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 34 / 67

Example
Z∞ Z∞ Z∞
dx 2 x2
Examine for convergence the integrals sin x dx, , dx.
0 −∞ 1 + x2 2 x4−1
Solution : By definition
Z∞ ZX

aft
sin x dx = lim sin x dx = lim [− cos x]X
0 = lim (1 − cos X)
0 X→∞ 0 X→∞ X→∞

Thus the improper integral does not exist since cos X has no limit when X → ∞.
Z∞ ZX
dx dx
= lim
−∞ 1 + x2 X→∞ Y 1 + x2
Y→−∞
  π  π
= lim tan−1 X − tan−1 Y = − − =π
X→∞ 2 2

Z∞
2 x2
x4 − 1
Dr Y→−∞

Thus, the integral exists and is equal to π.

dx = lim
X→∞
ZX
2 x2
x4 − 1
dx
2 2
 
1 X−1 1
= limtan−1 X − tan−1 2 + log + log 3
X→∞ 2 X+1 2
π −1 1
= − tan 2 + log 3
2 2
Thus, the integral converges.
Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 35 / 67
Example
Z∞ Z∞
dx 2
Examine for convergence the integrals , x3 e−x dx.
−∞ (x2 + 1)2 0
Solution : By definition
Z∞ Z0 Z∞
dx dx dx

aft
= +
−∞ (x2 + 1)2 −∞ (x2 + 1)2 (x2 + 1)2
0
Z∞ ZX
dx dx
=2 2 2
= 2 lim
0 (x + 1) X→∞ 0 (x2 + 1)2
 
X
= 2 lim tan−1 X + by putting x = tan θ
X→∞ 1 + X2
π
=
2
Thus, the integral converges.
Dr
Z∞

0
2
x3 e−x dx = lim
X→∞ 0
ZX
x3 e−x dx
2

 
1 1 2
= lim − (X2 + 1) e−X
X→∞ 2 2
1
=
2
Thus, the integral converges.

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 36 / 67

Comparison Tests for Convergence at ∞ (Integrand retaining its sign)
By definition
Z∞ ZX
f dx = lim f dx
a X→∞ a

where f is assumed to be bounded and integrable in [a, X] for every X > a.

aft
As in the case of unbounded functions with finite limits of integration, there is no loss of generality
in supposing that the integrand f is positive in [a, X], for, if negative it Zcan be replaced by (−f ).
∞
Theorem : A necessary and sufficient condition for the convergence of f dx, where f is positive
a
in [a, X] is that there exists a positive number M, independent of X, such that
ZX
f dx < M, for every X > a.
a

Proof : The integral

Z∞

a
Dr
f dx is said to be convergent if lim

Since f is positive in [a, X], the positive function of X,

X→∞ a
ZX
ZX
f dx exists finitely.

f dx, is monotone increasing as X increases

a
and will therefore tend to a finite limit if and only if it is bounded above, i.e., there exists a positive
number M, independent of X such that
ZX
f dx < M, for every X > a.
a

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 37 / 67

Comparison Test I (Comparison of two integrals)

If f and g be two positive functions such that f (x) 6 g(x), for all x in [a, X], then
Z∞ Z∞
(i) f dx converges, if g dx converges, and

aft
a a
Z∞ Z∞
(ii) g dx diverges, if f dx diverges.
a a
Proof : Let f and g be both bounded and integrable in [a, X], X > a.
Since f and g are positive and
f (x) 6 g(x), ∀ x ∈ [a, X]

ZX ZX

(i) Let
Z∞

a
Dr ∴
a
f dx 6
a
g dx

g dx be convergent, so that there exists a positive number M such that

(16)

ZX
g dx < M, for X > a
a

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 38 / 67

 Thus, from inequality (16)
ZX

aft
f dx < M, X > a.
a
Z∞
Hence f dx is convergent.
a
Z∞ ZX
(ii) If f dx is divergent then the positive function f dx is not bounded above and therefore
a ZX a

in view of (16), g dx is also not bounded above.

Hence,
Z∞

a
a Dr
g dx diverges.

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 39 / 67

Comparison Test II (Limit form)

f (x)
If f and g are positive in [a, X] and lim = l, where l is a non-zero finite number, then the two
x→∞ g(x)
Z∞ Z∞ Z∞
f
integrals f dx and g dx converge or diverge together. Also if → 0 and g dx converges

aft
Z∞ a a Z∞ Z ∞g a
f
then f dx converges, and if → ∞ and g dx diverges then f dx diverges.
a g a a
Proof : Evidently l > 0.
Let  be a positive number such that l −  > 0.
f (x)
Since lim = l, therefore there exists a number k (> a), however large, such that for all x > k,
x→∞ g(x)

or
Dr f (x)
g(x)
−l <

(l − ) g(x) < f (x) < (l + ) g(x) (17)

Now,
(l − ) g(x) < f (x), ∀x > k > a

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 40 / 67

 Z∞ Z∞ Z∞
So if f dx converges then by comparison test I, g dx and therefore g dx converges at ∞.
k k a
Again from inequality (17),
f (x) < (l + ) g(x), ∀ x > k > a
Z∞ Z∞ Z∞
So that if f dx diverges then by comparison test, g dx and therefore, g dx diverges at ∞.
k k a

aft
f
When → 0, we can find k so that
g

f (x)
< , ∀x > k
g(x)

i.e.,
f (x) <  g(x), ∀x > k
Z∞ Z∞
So, if

When
a
f
g
g dx converges thenDr a
f dx also converges.

→ ∞, we can find numbers k, M such that

f (x)
> M or f (x) > M g(x), ∀x > k
g(x)
Z∞ Z∞
Hence, if g dx diverges then f dx also diverges.
a a

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 41 / 67

A Useful Comparison Integral

Z∞
C
Show that the improper integral dx, a > 0, where C is a positive constant, converges if and
xn

aft
a
only if n > 1.
Proof : We have 
 X
ZX 
C log a , n=1
C
dx =
xn 
 
 1 1 1
a  − n−1 , n,1
1−n X n−1 a
and therefore 


Z∞
C
Z∞

a
Dr
C
xn
dx = lim
X→∞
ZX

a
C
xn
dx =

+∞,



C
(n − 1) an−1
,
n61

n>1

Thus dx converges if and only if n > 1.

a xn

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 42 / 67

Observations

Z∞
i) If f is positive in [a, x], then the integral f dx converges, if there exists a positive number n
a
greater than 1 and a fixed positive number M such that

aft
M
f (x) 6
xn
for every x > a.
Also the integral diverges if there exists a positive number M such that

M
f (x) >
xn
for every x > a.
Dr
ii) If lim [xn f (x)] exists and is non-zero finite, then the integral
x→∞
Z∞
f dx
a

converges if and only if n > 1.

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 43 / 67

Example

Z∞ Z∞
dx x2 dx
Examine the convergence of √ , √ .
1 x x2 + 1 0 x5 + 1
1 1
Solution : Let f (x) = √ , (behaves like x−2 at ∞) and g(x) = 2 ,

aft
x x2 + 1 x
so that
f (x) x2
= lim √
lim = 1 (non-zero finite)
g(x) x→∞x→∞ x x2 + 1
Z∞ Z∞
Hence, the two integrals f dx and g dx behave alike.
Z∞ 1 Z∞ 1
dx dx
As 2
converges, therefore √ also converges.
1 x 1 x x2 + 1

Let f (x) = √
x2

x5 + 1
 

f (x)
Dr
, ∼ x−1/2 and g(x) = √ .

= √
x5/2
1

= √
x

1
→ 1 as x → ∞
g(x) x5 + 1 1 + x−5
Z∞ Z∞ Z∞
dx x2 dx
As g dx = √ diverges, thus by comparison test, √ also diverges.
0 0 x 0 x5 + 1

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 44 / 67

Example
Z∞ Z∞
2 log x
Examine the convergence of e−x dx, dx.
0 x2 1
Solution : 0 is not a point of infinite discontinuity and so we have to examine convergence at ∞
only. Z ∞

aft
2
Let us consider the integral e−x dx.
1
We know
2
ex > x2 , for all real x
2 1
∴ e−x <
x2
Z∞ Z∞ Z∞
1 −x2 2
As 2
dx converges at ∞, the integral e dx and therefore the integral e−x dx con-
1 x 1 0
verges.
Here
Dr log x
x
log x
x3/2 2 = 1/2 → 0 as x → ∞
x
and Z∞
dx
1 x3/2
converges. Z∞ Z∞
dx log x
Hence by comparison with , the integral dx also converges.
1 x3/2 1 x2

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 45 / 67

Example
Z∞ Z∞
sin2 x
Examine the convergence of xn e−x dx, dx.
1 0 x2
Solution : We have
Z∞
xn+2

aft
dx
x2 xn e−x = → 0 as x → ∞, for all n and converges.
ex 1 x2
Z∞
Hence by comparison test, xn e−x dx converges.
1
Here 0 is not a point of infinite discontinuity, so if we put
Z∞ Z1 Z∞
sin2 x sin2 x sin2 x
dx = dx + dx
0 x2 0 x2 1 x2
Dr
the first integral on the right being proper, we test the second,
Z∞
sin2 x
x2
dx for convergence at ∞.
1
Z∞ Z∞
sin2 x 1 dx sin2 x
Now 6 2 and is convergent, therefore the integral dx is also convergent.
x2 x x2 x2
Z∞ 1 2 1
sin x
Hence, the integral dx is convergent.
0 x2

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 46 / 67

Example
Z∞ Z∞
x tan−1 x dx
Test for convergence the integrals dx, .
0 (1 + x4 )1/3 e2 x log log x
Solution : Let

x tan−1 x tan−1 x 1
(∼ x−1/3 at ∞) and g(x) = 1/3

aft
f (x) = = 1/3
(1 + x4 )1/3 x (1 + x−4 )1/3 x

so that
f (x) tan−1 x π
= → as x → ∞
g(x) (1 + x−4 )1/3 2
Z∞ Z∞
Hence, f dx and g dx behave alike.
Z ∞1 1 Z∞
dx x tan−1 x
Since
1 x
1/3
diverges,
t
Dr
therefore

Putting log x = t or x = e , we get

0 (1 + x )

Z∞
4 1/3

dx
dx also diverges.

=
Z∞
dt
e2 x log log x 2 log t

which diverges by the comparison with the divergent integral

Z∞
dt
m
, m 6 1.
k t

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 47 / 67

Gamma Function

Z∞
The integral xm−1 e−x dx is convergent if and only if m > 0.
0
Solution : Let

aft
e−x
f (x) = xm−1 e−x = .
x1−m
The integrand f has infinite discontinuity at 0 if m < 1. So we have to examine convergence at 0
and ∞ both.
Putting
Z∞ Z1 Z∞
xm−1 e−x dx = xm−1 e−x dx + xm−1 e−x dx
0 0 1
we test the two integrals on the right for convergence at 0 and ∞ respectively.
Dr
Convergence at 0 : m < 1.
Let
1
g(x) = 1−m
x
so that
f (x)
= e−x → 1 as x → 0
g(x)

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 48 / 67

Also Z1 Z1
dx
g dx =
0 0 x1−m
converges if and only if 1 − m < 1, i.e., m > 0.
Z1

aft
Hence xm−1 e−x dx converges if and only if m > 0.
0
Convergence at ∞
Let
1
g(x) =
x2
so that
f (x) xm+1
= x → 0 as x → ∞, for all m

As

Hence
Z∞

1
dx
Zx
∞
2
converges, therefore
Dr
g(x)
Z∞

1
e

xm−1 e−x dx also converges for all m.

xm−1 e−x dx is convergent if and only if m > 0.

0 Z∞
This integral xm−1 e−x dx is called Gamma function and is denoted by Γ (m).
0

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 49 / 67

Example

Examine for convergence Z∞  

1 1 dx
− .
0 x sinh x x

aft
Solution : Let  
1 1 1
f (x) = − .
x sinh x x
1
Using L’Hospital rule, we find that f (x) → as x → 0. Therefore 0 is not a point of infinite
6
discontinuity of f .
To examine the convergence at ∞, we put

f (x) =

so that f (x) behaves as



1
x2
Dr
1
x
−
1


sinh x x

at ∞.
1 1
= 2 −
x
2 e−x
x (1 − e−2x )

Z∞ Z∞  
dx 1 1 dx
By comparison with the convergent integral 2
we can show that − and
Z∞   1 x 1 x sinh x x
1 1 dx
therefore − is also convergent.
0 x sinh x x

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 50 / 67

Example

Test for the convergence of the integral

Z1
1 q
 
xp log dx.

aft
0 x

1
Solution : Making the substitution log = t or x = e−t , we get
x
Z1 Z∞
1 q
 
xp log dx = tq e−(p+1) t dt
0 x 0

The integrand of the last integral has a point of infinite discontinuity 0. So, we put
Z∞

0
Dr
tq e−(p+1) t dt =
Z1

0
tq e−(p+1) t dt +
Z∞

and examine the integrals on the right for convergence at 0 and ∞ respectively.
tq e−(p+1) t dt

Convergence at 0 : Let
e−(p+1) t
f (t) = tq e−(p+1) t = .
t−q

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 51 / 67

 Z1
dt
By comparison with , we find that
0 t−q
Z1
tq e−(p+1) t dt
0

aft
converges at 0 if only if −q < 1 or q > −1, for all values of p.
Convergence at ∞ :
Let
tq 1
f (t) = tq e−(p+1) t = (p+1) q and g(t) = 2
e t
so that
f (t) tq+2
= (p+1) t → 0, as t → ∞ for all q and p + 1 > 0.
g(t) e
Z∞ Dr
dt
Also 2
converges.
1Z t
∞
Hence tq e−(p+1) t dt converges for all q and only if p + 1 > 0.
1
Hence Z∞ Z1
1 q
 
tq e−(p+1) t dt = xp log dx
1 0 x
is convergent for p > −1, q > −1 and divergent for all other values of p and q.

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 52 / 67

General Test for Convergence at ∞ (Integrand may change sign)
Z∞
Caychy’s Test : The integral f dx converges at ∞ if and only if for every  > 0 there corresponds
a
a positive number X0 such that
Z X2

aft
f dx < , for all X1 , X2 > X0
X1

Z X2
[It implies that f dx → 0 as X1 , X2 → ∞]
X1
Z∞ ZX
Proof : The improper integral f dx exists if lim f dx exists finitely.
a X→∞ a
Let ZX

a function of X.
Dr F(X) =
a
f dx,

According to Cauchy’s criterion for finite limits F(X) tends to a finite limit as X → ∞ if and only
if for every  > 0 there corresponds X0 such that for all X1 , X2 > X0 ,

|F(X1 ) − F(X2 )| < 

i.e.,
Z X1 Z X2 Z X2
f dx − f dx <  ⇒ f dx < .
a a X1

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 53 / 67

Example
Z∞
sin x
Show that dx is convergent.
0 x
sin x
Solution : Here 0 is not a point of infinite discontinuity, for → 1 as x → 0. So, let us put
x
Z∞ Z1 Z∞

aft
sin x sin x sin x
dx = dx + dx
0 x 0 x 1 x
Z∞
sin x
and test dx for convergence at ∞.
0 x
2
For any  > 0, let X1 , X2 be two numbers, both greater than .

Now Z X2 Z X2
sin x h cos x iX2 cos x

so that
Dr
Z X2
sin x
dx 6
X1

cos X1
x

−
dx = −

cos X2
x X1

+
Z X2
−

cos x
2
dx
X1 x2
dx

X1 x X1 X2 X1 x
Z X2
1 1 dx 
6 + + 2
<2· =
X1 X2 X1 x 2
Z∞ Z∞
sin x sin x
Then by Cauchy’s test dx and consequently dx is convergent.
1 x 0 x

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 54 / 67

Absolute Convergence

aft
Z∞ Z∞
The improper integral f dx is said to be absolutely convergent if |f | dx is convergent.
a a

The following theorem can be proved with the help of Cauchy’s text.
Z∞ Z∞
Theorem : The improper integral f dx exists, if |f | dx exists.
a a
Dr
The converse, however, is not true. Integrals which are convergent but not absolutely are called
conditionally convergent integrals.

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 55 / 67

Example

Z∞
sin x
Show that dx converges absolutely if p > 1.
xp

aft
1
Solution : Now
sin x | sin x| 1
= 6 p, ∀x > 1
xp xp x
Also Z∞
dx
converges only if p > 1.
1 xp
Z∞
sin x

Hence
Dr
∴

Z∞

1
sin x
xp
1 xp
dx converges if p > 1.

dx converges absolutely if p > 1.

For p = 1 it is not absolutely convergent.

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 56 / 67

 aft
Integrand as a Product of Functions
(Convergence at ∞)
Dr
Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 57 / 67
A Test for Absolute Convergence
Z∞
A function φ is bounded in [a, ∞) and integrable in [a, X] where X is a number > a. If f dx is
Z∞ a

absolutely convergent at ∞, then f φ dx is also absolutely convergent at ∞.

a
Proof : Since φ is bounded in [a, ∞), there exists a positive K such that

aft
φ(x) 6 K, ∀x > a (18)
Z∞
Again, since |f | is positive and |f | dx is convergent, a number M exists such that
a
ZX
|f | dx 6 M, for all X > a (19)
a

Using inequality (18), we have

⇒
Dr ZX
|f φ| 6 K |f |, ∀ x > a

|f φ| dx 6 K
ZX
|f | dx 6 K M, for all X > a
a a
ZX
which implies that the positive function |f φ| dx is bounded above by K M, for X > a. Hence
Z∞ Z∞ a

|f φ| dx converges ⇒ f φ dx converges absolutely.

a a

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 58 / 67

Example

aft
Discuss the convergence of the integral

Z∞ 

1
 2, x is a rational number
x
f (x) dx, where f (x) =


1 − 1 , x is an irrational number
x2
Solution : Since Z∞ Z∞
1
Dr 1
|f (x)| dx =
1 x2
dx

is convergent, and every absolutely convergent integral is convergent. Therefore the given integral
is convergent.

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 59 / 67

 aft
Tests for Convergence

Dr
Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 60 / 67
Abel’s Test

Z∞ Z∞
If φ is bounded and monotonic in [a, ∞), and f dx is convergent at ∞, then f φ dx is
a a
convergent at ∞.

aft
or
An infinite integral which converges (not necessarily absolutely) will remain convergent after the
insertion of a factor which is bounded and monotonic.
Proof : Since φ is monotonic in [a, ∞), it is integrable in [a, X], for all X > a. Also, since f is
integrable in [a, X], we have by Second Mean Value Theorem
Z X2 Zξ Z X2
f φ dx = φ (X1 ) f dx + φ (X2 ) f dx for a < X1 6 ξ 6 X2 (20)
X1 X1 ξ
Dr
Let  > 0 be arbitrary. Since φ is bounded in [a, ∞), a positive number K exists such that

|φ(x)| 6 K, ∀x > a

In particular,

|φ (X1 )| 6 K, |φ (X2 )| 6 K (21)

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 61 / 67

 Z∞
Again, since f dx is convergent, a number X0 exists such that
a

Z X2

f dx < , for all X1 , X2 > X0 (22)
X1 2K

aft
Let the numbers X1 , X2 in (20) be > X0 so that the number ξ which lies between X1 and X2 is also
> X0 .
Hence from (22)
Z X2 Zξ Z X2
f φ dx = φ (X1 ) f dx + φ (X2 ) f dx
X1 X1 ξ
Zξ Z X2
6 |φ(X1 )| · f dx + |φ(X2 )| · f dx

Hence, by Cauchy’s test,

Dr <K·

2K
+K·

2K
Z∞
=
X1 ξ

f φ dx
a
is convergent at ∞.

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 62 / 67

Dirichlet’s Test

ZX
If φ is bounded and monotonic in [a, ∞) and tends to 0 as x → ∞, and f dx is bounded for
Z∞ a

f φ dx is convergent at ∞.

aft
X > a, then
a
or
An infinite integral which oscillates finitely becomes convergent after the insertion of a bounded
monotonic factor which tends to zero as a limit.
Proof : Since φ is monotonic, it is integrable in [a, X] for all X > a.
Also, since f is integrable in [a, X]. Therefore by Second Mean Value Theorem,
Z X2 Zξ Z X2
f φ dx = φ (X1 ) f dx + φ (X2 ) f dx for a < X1 6 ξ 6 X2 (23)

Again, since
X1

ZX

a
Dr X1 ξ

f dx is bounded when X > a, there exists a number K > 0 such that

ZX
f dx 6 K, ∀x > a
a

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 63 / 67

 Zξ Zξ Z X1 Zξ Z X1
f dx = f dx − f dx 6 f dx + f dx
X1 a a X1 a

6 K + K = 2 K, for X1 , ξ > a
Similarly
Z X2
f dx 6 2 K, for X2 , ξ > a

aft
ξ

Let  > 0 be arbitrary.

Since φ → 0 as x → ∞, a positive number X0 exists such that
 
|φ (X1 )| < , |φ (X2 )| < where X2 > X1 > X0
4K 4K
Let the numbers X1 , X2 in (23) be > X0 , so that we get
Z X2 Zξ Z X2

X1
Dr
f φ dx = φ (X1 ) f dx + φ (X2 )

6 |φ (X1 )| ·
X1
Zξ

X1
f dx
ξ

f dx + |φ (X2 )| ·
Z X2

ξ
f dx

 
< · 2K + · 2K = 
4K 4K
Hence, by Cauchy’s test, Z∞
f φ dx
a
is convergent at ∞.
Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 64 / 67
Example

Z∞
sin x
The integral dx is convergent for p > 0.
1 xp
Solution : Let
f (x) = sin x

aft
and
1
φ(x) = .
xp
Here
ZX ZX
f dx = sin x dx = |cos 1 − cos X|
1 1

6 |cos 1| + |cos X|

1
Dr 6 2, for 1 6 X < ∞.

Also φ(x) = p is a monotone decreasing function tending to 0 as x → ∞ for p > 0.

x
Therefore, by Dirichlet’s test, Z∞ Z∞
sin x
f φ dx = dx
1 1 xp
converges for p > 0.

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 65 / 67

We already shown that Z∞
sin x
dx
1 xp
converges absolutely for p > 1.

aft
Thus we conclude that Z∞
sin x
dx
1 xp
converges absolutely for p > 1, but only conditionally for 0 < p 6 1.

The integral Z∞
sin x log x
dx
1 x
is convergent.
Let

Z∞
Drf (x) = sin x, φ(x) =
log x
x
.

Now sin x dx is bounded above by 2, and φ is monotone decreasing to 0 as x → ∞.

e

Hence, the given integral converges by Dirichlet’s test.

Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 66 / 67

 aft
Dr
Sharad Bhagwanrao Kadam Improper Integrals April 24, 2021 67 / 67