IMPROPER INTEGRALS

(ADDITIONAL NOTES TO STEWART §7.8)

Definition 1

Suppose that a ∈ R and f is continuous on [a, ∞). Then

Z ∞ Z t
f (x) dx = lim f (x) dx
a t→∞ a

provided that this limit exists.

Suppose that b ∈ R and f is continuous on (−∞, b]. Then
Z b Z b
f (x) dx = lim f (x) dx
−∞ t→−∞ t

provided
R∞ that this
R b limit exists.
a
f (x) dx and −∞ f (x) dx are called improper integrals of Type 1. An improper integral of Type
1 is called convergent if the limit involved exists, and it is called divergent if the limit involved does
not exist.
R∞ Ra
If both a f (x) dx and −∞ f (x) dx are convergent (a any number), then we define
Z ∞ Z a Z ∞
f (x) dx = f (x) dx + f (x) dx.
−∞ −∞ a
R∞
−∞
f (x) dx is also called an improper integral of Type 1.

Definition 2

Suppose that f is continuous but unbounded on [a, b). Then

Z b Z t
f (x) dx = lim f (x) dx
a t→b− a

provided that this limit exists.

Suppose that f is continuous but unbounded on (a, b]. Then
Z b Z b
f (x) dx = lim f (x) dx
a t→a+ t

provided that this limit exists.

These integrals are called improper integrals of Type 2. An improper integral of Type 2 is called
convergent if the limit involved exists, and it is called divergent if the limit involved does not exist.
Suppose that a < c < b, f is discontinuous at c and f is continuous but unbounded on [a, c) and on
Rc Rb
(c, b]. If both a f (x) dx and c f (x) dx are convergent, then we define
Z b Z c Z b
f (x) dx = f (x) dx + f (x) dx.
a a c

Such an integral is also called an improper integral of Type 2.

1
Definition 3

Suppose that a ∈ R, f is continuous

R c on (a, ∞) and for any cR ∞ > a we have that f is unbounded on
(a, c]. If the improper integrals a f (x) dx (of Type 2) and c f (x) dx (of Type 1) converge, then
we define Z ∞ Z c Z ∞
f (x) dx = f (x) dx + f (x) dx.
a a c
Such an integral is called an improper integral of Type 3.

Lemma 4
R∞
1. RSuppose that a < b. If f is continuous on [a, ∞), then a f (x) dx is convergent if and only if
∞
b
f (x) dx is convergent.
R∞ R∞
2. RIf both a f (x) dx and a g(x) dx converge and α and β are numbers, then
∞
a
[αf (x) + βg(x)] dx converges as well.

Similar results also hold for improper integrals of Type 2.

Lemma 5
R∞ 1
Suppose that a < b. Then b (x−a)p
dx is convergent for p > 1 and divergent for p ≤ 1.

Lemma 6
R∞
Suppose that a ∈ R. Then a
e−px dx is convergent for p > 0 and divergent for p ≤ 0.

Lemma 7
Rb 1
Rb 1
Suppose that a < b. Then a (x−a)p
dx and a (x−b)p
dx are convergent if p < 1 and divergent if p ≥ 1.

Before we can obtain further results, we must first study the concept of a smallest upper bound.

Definition 8

1. A set A of real numbers is bounded above if there exists a real number M such that x ≤ M for
all x ∈ A. We call M an upper bound for A.

2. A set A of real numbers is bounded below if there exists a real number m such that x ≥ m for
all x ∈ A. We call m a lower bound for A.

3. A set A ⊂ R is bounded if A is both bounded below and bounded above.

2
 4. If A ⊂ R is not bounded, we say that A is unbounded.

Definition 9

If a set A ⊂ R has a smallest upper bound S, then we call S the supremum or sup of A and write
S = sup A. This means:
1. S is an upper bound for A.

2. If T is any upper bound for A, then S ≤ T .

Theorem 10

Let A ⊂ R. Then sup A = S if and only if the following two conditions hold:
1. S is an upper bound for A.

2. For each  > 0 there exists an x ∈ A such that x > S − .

Proof

If sup A = S, then (1) holds, since S is an upper bound (the smallest one) for A. For any  > 0 we
have that S −  < S, and hence S −  is smaller than the smallest upper bound S, so that S −  is
not an upper bound for A. Hence there exists at least one x ∈ A such that x > S − . So (2) holds.
Conversely, suppose that (1) and (2) hold. Then S is an upper bound for A. It follows from (2) that
each number smaller than S is not an upper bound for A. Hence S is the smallest upper bound for
A, i.e. S = sup A. 

Definition 11

If a set A ⊂ R has a greatest lower bound I, then we call I the infimum or inf of A and write
I = inf A. This means:
1. I is a lower bound for A.

2. If J is any lower bound for A, then I ≥ J.

Theorem 12

Let A ⊂ R. Then inf A = I if and only if the following two conditions hold:
1. I is a lower bound for A.

2. For each  > 0 there exists an x ∈ A such that x < I + .

The proof of Theorem 12 is similar to that of Theorem 10.

Finally, we formulate the following axiom:

3
The Completeness Property of R

Each non-empty set of real numbers which is bounded above has a smallest upper bound.

Corollary

Each non-empty set of real numbers which is bounded below has a greatest lower bound.

We are now ready to prove the following important theorem:

Theorem 13

Suppose that f is increasing on [a, ∞). Then lim f (x) exists if and only if f is bounded above on
x→∞
[a, ∞), in which case
lim f (x) = sup{f (x) : x ∈ [a, ∞)}.
x→∞

Proof

Suppose that f is increasing on [a, ∞).

We first prove that if lim f (x) exists, then f is bounded above on [a, ∞) (so that sup{f (x) : x ∈
x→∞
[a, ∞)} exists):
So suppose that lim f (x) exists, and let L = lim f (x). Then there exists a number N ≥ a so that
x→∞ x→∞
|f (x)−L| < 1 for all x ≥ N (take  = 1). Hence −1 < f (x)−L < 1, so that L−1 < f (x) < L+1, for
all x ≥ N . In particular, f (N ) < L + 1. Since f is increasing on [a, ∞), we have that f (x) ≤ f (N ),
so that f (x) < L + 1, for all x ∈ [a, N ]. It follows that f (x) < L + 1 for all x ∈ [a, ∞), and hence f
is bounded above on [a, ∞).
Next we prove that if f is bounded above on [a, ∞), then lim f (x) exists and lim f (x) = sup{f (x) :
x→∞ x→∞
x ∈ [a, ∞)}:
So suppose that f is bounded above on [a, ∞). Let S = sup{f (x) : x ∈ [a, ∞)}. Then there exists
for each  > 0 a y ∈ [a, ∞) such that S −  < f (y). Since f is increasing on [a, ∞), we have that
f (y) ≤ f (x), so that S −  < f (x), for all x ≥ y. Also, f (x) ≤ S < S +  for all x ∈ [a, ∞), so that
S −  < f (x) < S + , i.e. |S − f (x)| < , for all x ≥ y. It follows from this that lim f (x) exists and
x→∞
lim f (x) = S.
x→∞
The result follows. 

Using the above theorem we can extend the theory of improper integrals further:

4
The comparison test for improper integrals of Type 1

Suppose that a ∈ R and f and g are continuous functions on [a, ∞) such that 0 ≤ g(x) ≤ f (x) for
all x ≥ a.
R∞ R∞
1. If a f (x) dx is convergent, then a g(x) dx is convergent.
R∞ R∞
2. If a g(x) dx is divergent, then a f (x) dx is divergent.

Proof
Rt Rt
1. Let G(t) = a g(x) dx and F (t) = a f (x) dx. Since g(x) ≤ f (x) for all x ≥ a, it follows that
G(t) ≤ F (t) for all t ≥ a. The fact that f (x) ≥ 0 and g(x) ≥ 0 for all x ≥ a implies that the
functions F and G are increasing on [a, ∞).
R∞ R∞
Now suppose that a f (x) dx is convergent. Then lim F (t) exists (with lim F (t)= a f (x) dx).
t→∞ t→∞
Since F is increasing on [a, ∞) and lim F (t) exists, it follows from Theorem 13 that {F (t) :
t→∞
t ∈ [a, ∞)} is bounded above. Since G(t) ≤ F (t) for all t ≥ a, we have that {G(t) : t ∈ [a, ∞)}
is bounded above as well. Together with theR fact that G is increasing on [a, ∞), it follows from
∞
Theorem 13 that lim G(t) exists, i.e. that a g(x) dx converges.
t→∞

2. This is the contrapositive of (1).

The comparison test for improper integrals of Type 2

Suppose that a < b, f and g are continuous but unbounded on (a, b] and 0 ≤ g(x) ≤ f (x) for all
x ∈ (a, b].
Rb Rb
1. If a f (x) dx is convergent, then a g(x) dx is convergent.
Rb Rb
2. If a g(x) dx is divergent, then a f (x) dx is divergent.

The above theorem also holds if the interval (a, b] is replaced by [a, b). The proofs are similar to the
proof of the comparison test for improper integrals of Type 1.

The quotient test for improper integrals of Type 1

Suppose that a ∈ R and f and g are continuous functions on [a, ∞) such that f (x) ≥ 0 and g(x) ≥ 0
for all x ≥ a.
f (x) R∞ R∞
1. If lim = 0 and a g(x) dx is convergent, then a f (x) dx is convergent.
x→∞ g(x)

f (x) R∞ R∞
2. If lim = A > 0, then a g(x) dx is convergent if and only if a f (x) dx is convergent.
x→∞ g(x)

f (x) R∞ R∞
3. If lim = ∞ and a g(x) dx is divergent, then a f (x) dx is divergent.
x→∞ g(x)

5
Proof
f (x) R∞ f (x)
1. Suppose that lim = 0 and a g(x) dx is convergent. Since lim = 0, there exists
x→∞ g(x) x→∞ g(x)

a number N — and we may take N ≥ a — such that fg(x) (x)

− 0 < 1 (take  = 1), and hence
f (x) R∞
g(x)
< 1, or 0 ≤ f (x) < g(x), for all x ≥ N . By Lemma 4 N g(x) dx is convergent, and
R∞
hence it followsRfrom the comparison test that N f (x) dx is convergent. Hence it follows from
∞
Lemma 4 that a f (x) dx is convergent.

2. Suppose that limx→∞ fg(x)

(x)
= A > 0. Let m and M be positive numbers such that m < A < M .
f (x)
Since lim = A, it follows that there exists a number N ≥ a such that m ≤ fg(x)
(x)
≤ M for
x→∞ g(x)
R∞
all x ≥ N . Hence 0 ≤R ∞mg(x) ≤ f (x) ≤ M g(x) for all Rx∞≥ N . Now suppose that a g(x) dx
is convergent. Then N g(x) dx is convergent, R ∞so that N M g(x) dx is convergent,
R ∞by Lemma
4. By the comparison test it followsR that N f (x) dx is convergent, so that a
f (x) dx is
∞ R∞
R ∞ a f (x) dx is convergent. Then N f (x) dx is convergent,
convergent. Conversely, suppose that
R ∞ by the comparison test, N mg(x) dx is convergent. Hence it follows from Lemma 4
so that,
that a g(x) dx is convergent.

f (x) R∞ f (x)
3. Suppose that lim = ∞ and a g(x) dx is divergent. Since lim = ∞, there exists
x→∞ g(x) x→∞ g(x)

a number N such that fg(x)

(x)
> 1, i.e. 0 ≤ g(x) < f (x), for all x ≥ N . It follows from Lemma 4
R∞ R∞
that NR g(x) dx is divergent and hence, by the comparison test, that N f (x) dx is divergent.
∞
Hence a f (x) dx is divergent. 

The quotient test for improper integrals of Type 2

Suppose that a < b, f and g are continuous but unbounded on (a, b] and f (x) ≥ 0 and g(x) ≥ 0 for
all x ∈ (a, b].
f (x) Rb Rb
1. If lim = 0 and a g(x) dx is convergent, then a f (x) dx is convergent.
x→a+ g(x)

f (x) Rb Rb
2. If lim = A > 0, then a g(x) dx is convergent if and only if a f (x) dx is convergent.
x→a+ g(x)

f (x) Rb Rb
3. If lim = ∞ and a g(x) dx is divergent, then a f (x) dx is divergent.
x→a+ g(x)

A similar theorem as the above also holds if the interval (a, b] is replaced by [a, b). The proofs are
similar to the proof of the quotient test for improper integrals of Type 1.

We conclude with an important concept:

Definition 14
R∞ R∞
Suppose that f is continuous on [a, ∞). Then a
f (x) dx is absolutely convergent if a
|f (x)| dx is

6
convergent.
R∞ R∞ R∞
a
f (x) dx is conditionally convergent if a f (x) dx is convergent and a |f (x)| dx is divergent.

Similar definitions also hold for improper integrals of Type 2.

Theorem 15
R∞ R∞
If a
f (x) dx is absolutely convergent, then a
f (x) dx is convergent.

Proof
R∞
For x ≥ a we have that 0 ≤R f (x)+|f (x)| ≤ 2|f (x)|. Since 2 a |f (x)| dx is convergent, it follows using
∞
the comparison test that a [f (x) + |f (x)|] dx is convergent. Since f (x) = [f (x) + |f (x)|] − |f (x)|,
the result follows from Lemma 4.