Page 1 of 125

MA1201 Calculus and Basic Linear Algebra II

Chapter 2
Integration Technique
 Page 2 of 125

In this Chapter, we shall discuss some important technique in evaluating integrals:

 Method of Substitution (pp. 3=37)

 Integration by parts & Reduction Formula (pp. 38-89)

 Integration of rational functions – Method of partial fractions (pp.90-110)

 Page 3 of 125

Method of substitution
The principle of this method is to “transform” some complicated integrals into some
simpler integrals by introducing a new variable 𝑦.

3𝑥 2 1
∫ 3 𝑑𝑥 𝑦 = 𝑥3 + 1 ∫ 𝑑𝑦 = ln |𝑦| + 𝐶
𝑥 +1 𝑦

(Difficult to integrate) (Easy to integrate)

Procedure of transformation
𝑑𝑦 1
Let 𝑦 = 𝑥 3 + 1, then = 3𝑥 2 ⇒ 𝑑𝑥 = 𝑑𝑦.
𝑑𝑥 3𝑥 2
3𝑥 2 3𝑥 2 1 1
∫ 3 𝑑𝑥 ∫ 3 ( 2 𝑑𝑦) ∫ 𝑑𝑦
𝑥 +1 𝑥 + 1 3𝑥 𝑦
Change 1 Express the
𝑑𝑥 to 𝑑𝑦
=∫ 3 𝑑𝑦 integrand in terms
𝑥 +1
of 𝑦
 Page 4 of 125

Example 1
Compute the integral

∫ 𝑥 2 cos(𝑥 3 + 1) 𝑑𝑥

Solution:
IDEA: Since the function cos(𝑥 3 + 1) is hard to integrate, it may be easier for
us to compute if we “transform” this function into cos 𝑦.

Step 1: Choose a substitution

We let 𝑦 = 𝑥 3 + 1.
Step 2: Transform the integral (𝒙 → 𝒚) and compute the transformed integral
The derivative can be computed as
𝑑𝑦 2
1
= 3𝑥 ⇒ 𝑑𝑥 = 2 𝑑𝑦,
𝑑𝑥 3𝑥
 Page 5 of 125

1
∫ 𝑥 2 cos(𝑥 3 + 1) 𝑑𝑥 = ∫ 𝑥 2 cos(𝑥 3 + 1) ( 2
𝑑𝑦)
3𝑥
1
= ∫ cos(𝑥 3 + 1) 𝑑𝑦
3
1 1
= ∫ cos 𝑦 𝑑𝑦 = sin 𝑦 + 𝐶
3 3
substitue 𝑦=𝑥 3 +1
1
=
⏞ sin(𝑥 3 + 1) + 𝐶.
3
Checking:
( ( 3 )) ( 3
𝑑 1 3
1 𝑑 sin 𝑥 + 1 𝑑 𝑥 + 1) 2 3
( sin(𝑥 + 1)) = = 𝑥 cos( 𝑥 + 1).
𝑑𝑥 3 3 𝑑 (𝑥 3 + 1) 𝑑𝑥
Reminder:
Remember to transform the answer into function of 𝑥 after integration!
 Page 6 of 125

Example 2
Compute the integral

∫ 𝑥 5 √𝑥 2 + 3 𝑑𝑥

Solution:
IDEA: Since the function √𝑥 2 + 3 is hard to integrate, it may be easier for us
to compute if we “transform” this function into √𝑦.

Step 1: Choose a substitution

We let 𝑦 = 𝑥 2 + 3.
Step 2: Transform the integral (𝒙 → 𝒚) and compute the transformed integral
𝑑𝑦 1
= 2𝑥 ⇒ 𝑑𝑥 = 𝑑𝑦,
𝑑𝑥 2𝑥
 Page 7 of 125

Then the integral becomes

1 1
∫ 𝑥 5 √𝑥 2 + 3𝑑𝑥 = ∫ 𝑥 5 √𝑥 2 + 3 ( 𝑑𝑦) = ∫ 𝑥 4 √𝑥 2 + 3 𝑑𝑦
2𝑥 2
𝑦=𝑥 2 +3
⇒𝑥 2 =𝑦−3
1 2
1 2
1
=
⏞ ∫(𝑦 − 3) √𝑦𝑑𝑦 = ∫(𝑦 − 6𝑦 + 9) (𝑦 2 ) 𝑑𝑦
2 2
1 5 3 1
= [∫ 𝑦 2 𝑑𝑦 − 6 ∫ 𝑦 2 𝑑𝑦 + 9 ∫ 𝑦 2 𝑑𝑦]
2
𝑥 𝑎+1
∫ 𝑥 𝑎 𝑑𝑥= +𝐶 5 3 1
𝑎+1 1 𝑦 2+1 𝑦 2+1 𝑦 2+1
=
⏞ [ −6 +9 ]+𝐶
2 5+1 3 1
+1 +1
2 2 2
𝑦=𝑥 2 +3
1 2 7 6 2 5
2
3
=
⏞ (𝑥 + 3)2 − (𝑥 + 3)2 + 3(𝑥 + 3)2 + 𝐶.
7 5
 Page 8 of 125

Remark of Example 2 (On the choice of substitution)

One may try to transform the integral by using the substitution 𝑦 = √𝑥 2 + 3. The
computation is more tedious although one can obtain the answer:
𝑑𝑦 1 𝑥
Note that = (2𝑥 ) = , then the integral becomes
𝑑𝑥 2√𝑥 2 +3 √𝑥 2 +3

√𝑥 2+3
∫ 𝑥 5 √𝑥 2 + 3𝑑𝑥 = ∫ 𝑥 5 √𝑥 2 + 3 𝑑𝑦 = ∫ 𝑥 4 (𝑥 2 + 3)𝑑𝑦
𝑥
7 5
𝑦 6𝑦
= ∫(𝑦 2 − 3)2 𝑦 2 𝑑𝑦 = ∫(𝑦 6 − 6𝑦 4 + 9𝑦 2 )𝑑𝑦 = − + 3𝑦 3 + 𝐶
7 5
1 2 7 6 2 5
2
3
= (𝑥 + 3)2 − (𝑥 + 3)2 + 3(𝑥 + 3)2 + 𝐶.
7 5
Therefore, one should choose some simple (but effective) substitution in order to
minimize the computation cost. Remember, “simple is the best”.
 Page 9 of 125

Method of substitution for definite integral

The procedure is similar to that of indefinite integral except the determination of
upper limits and lower limits of transformed integral since the original range of
integration is for the variable 𝑥 only. One needs to determine the range of
integration of new variable 𝑦 when using the method of substitution.

Example 3 (Definite integral)

Evaluate the integral

4 sin (
1
)
∫ 𝑥 𝑑𝑥 .
1 𝑥2

Solution
 Page 10 of 125

Step 1: Choose your substitution

1 1
We let 𝑦 = . (so that sin ( ) can turn into sin 𝑦)
𝑥 𝑥

Step 2: Transform the integral (𝒙 → 𝒚) and compute the integral

𝑑𝑦 𝑑 1 𝑑 −1 1
= ( )= 𝑥 = −𝑥 = − 2 ⇒ 𝑑𝑥 = −𝑥 2 𝑑𝑦
−2
𝑑𝑥 𝑑𝑥 𝑥 𝑑𝑥 𝑥
1 Determinant the upper limit
When 𝑥 = 4, then 𝑦 = ; When 𝑥 = 1, then 𝑦 = 1.
4 and lower limit of the new
1 1 1 integral
4 sin ( ) sin ( )
∫ 𝑥 𝑑𝑥 = ∫4 𝑥 (−𝑥 2 𝑑𝑦)
1 𝑥2 1 𝑥2
1
1
4 1
= − ∫ sin 𝑦 𝑑𝑦 = −[− cos 𝑦]14 = cos − cos 1.
1 4
 Page 11 of 125

Example 4
Evaluate
ln 2
𝑒 2𝑥
∫ 𝑑𝑥
0 √𝑒 𝑥 +1
Step 1: Choose your substitution

Choose 𝑦 = 𝑒 𝑥 + 1 (Question: Why not 𝑒 𝑥 or √𝑒 𝑥 + 1??)

Step 2: Transform the integral (𝒙 → 𝒚) and compute the integral

𝑑𝑦 𝑑 𝑥 𝑑𝑦
= (𝑒 + 1) = 𝑒 𝑥 ⇒ 𝑑𝑥 = 𝑥 .
𝑑𝑥 𝑑𝑥 𝑒
When 𝑥 = ln 2, 𝑦 = 𝑒 ln 2 + 1 = 3; When 𝑥 = 0, 𝑦 = 𝑒 0 + 1 = 2.
 Page 12 of 125

Then the integral becomes

ln 2 3 2𝑥
3 𝑥
𝑒 2𝑥 𝑑𝑦 𝑒𝑒
∫ 𝑑𝑥 = ∫ 𝑥
=∫ 𝑑𝑦
0
𝑥
√𝑒 + 1 2 √𝑦 𝑒 2 √𝑦

𝑦=𝑒 𝑥 +1 3
𝑦−1
=
⏞ ∫ 𝑑𝑦
2 √𝑦
3 3
𝑦 1 1
=∫ ( − ) 𝑑𝑦 = ∫ (√𝑦 − ) 𝑑𝑦
2 √𝑦 √𝑦 2 √𝑦
𝑥 𝑎+1 3 1 3
3 ∫ 𝑥 𝑎 𝑑𝑥= 𝑎+1
1
−
1 𝑦2 𝑦2 2 √2
=∫ (𝑦 2 − 𝑦 2 ) 𝑑𝑦 =
⏞ [ − ] = .
2
3 1 3
2 2 2
 Page 13 of 125

Example 5
Compute the integral
∫ tan 𝑥 𝑑𝑥.
Solution:
sin 𝑥
One has to rewrite the integrand as ∫ tan 𝑥 𝑑𝑥 = ∫ 𝑑𝑥.
cos 𝑥
Step 1: Choose your substitution
We choose the substitution 𝑦 = cos 𝑥 (so that the denominator becomes 𝑦 only).
Step 2: Transform the integral and compute it
𝑑𝑦
Note that = − sin 𝑥, then the integral becomes
𝑑𝑥
sin 𝑥 sin 𝑥 1 1
∫ 𝑑𝑥 = ∫ ( 𝑑𝑦) = − ∫ 𝑑𝑦 = − ln|𝑦| + 𝐶
cos 𝑥 cos 𝑥 − sin 𝑥 𝑦
= − ln|cos 𝑥 | + 𝐶 (or ln|sec 𝑥 | + 𝐶 )
 Page 14 of 125

Example 6
Evaluate
𝑥+1
∫ 2
𝑑𝑥.
𝑥 + 2𝑥 + 7
Solution:
Step 1: Choose your substitution: We let 𝑦 = 𝑥 2 + 2𝑥 + 7.
Step 2: Transform the integral (𝒙 → 𝒚) and compute the integral
𝑑𝑦 𝑑𝑦
= 2𝑥 + 2 ⇒ 𝑑𝑦 = (2𝑥 + 2)𝑑𝑥 ⇒ 𝑑𝑥 = .
𝑑𝑥 2𝑥 + 2
2𝑥+2
=2(𝑥+1)
𝑥+1 𝑥+1 𝑑𝑦 1 1 1
∫ 2 𝑑𝑥 = ∫ 2 =
⏞ ∫ 𝑑𝑦 = ∫ 𝑑𝑦
𝑥 + 2𝑥 + 7 𝑥 + 2𝑥 + 7 2𝑥 + 2 ( 2
2 𝑥 + 2𝑥 + 7) 2 𝑦
𝑦=𝑥 2 +2𝑥+7
1 1
= ln|𝑦| + 𝐶 =
⏞ ln|𝑥 2 + 2𝑥 + 7| + 𝐶.
2 2
 Page 15 of 125

Harder Examples
In some situations, one may not able to transform the integral into the desired form
using simple substitution. In this case, one has to either use algebra trick to turn the
integral into another form so that the simple substitution still works or use other
“more advanced” substitutions.
Example 7 (Related to Example 6)
Compute the integral
2𝑥 − 5
∫ 2 𝑑𝑥.
𝑥 − 6𝑥 + 10
Since this integral looks like the one in Example 5, one may try the substitution 𝑦 =
𝑑𝑦
𝑥 2 − 6𝑥 + 10 so that = 2𝑥 − 6. But we get
𝑑𝑥

2𝑥 − 5 2𝑥 − 5 𝑑𝑦 2𝑥 − 5
∫ 𝑑𝑥 = ∫ ( ) = ∫ 𝑑𝑦.
𝑥 2 − 6𝑥 + 10 𝑥 2 − 6𝑥 + 10 2𝑥 − 6 (2𝑥 − 6)𝑦

And we cannot move on since there is no cancellation between (2𝑥 − 5) and (2𝑥 − 6).
 Page 16 of 125

To settle this technical problem, one can modify the integral so that the “cancellation”
works.
Solution:
extra term
2𝑥 − 5 2𝑥 − 6 ⏞ 1
∫ 2 𝑑𝑥 = ∫ [( 2 )+ 2 ] 𝑑𝑥
𝑥 − 6𝑥 + 10 𝑥 − 6𝑥 + 10 𝑥 − 6𝑥 + 10

2𝑥 − 6 1
=∫ 2 𝑑𝑥 + ∫ 2 𝑑𝑥 … … … … . (∗)
𝑥 − 6𝑥 + 10 𝑥 − 6𝑥 + 10
2𝑥−6
For the 1st integral ∫ 2 𝑑𝑥, we let
𝑥 −6𝑥+10

2
𝑑𝑦 1
𝑦 = 𝑥 − 6𝑥 + 10 ⇒ = 2𝑥 − 6 ⇒ 𝑑𝑥 = 𝑑𝑦.
𝑑𝑥 2𝑥 − 6
2𝑥 − 6 2𝑥 − 6 1 1
∫ 2 𝑑𝑥 = ∫ 2 ( 𝑑𝑦) = ∫ 𝑑𝑦 = ln |𝑦| + 𝐶
𝑥 − 6𝑥 + 10 𝑥 − 6𝑥 + 10 2𝑥 − 6 𝑦
= ln|𝑥 2 − 6𝑥 + 10| + 𝐶.
 Page 17 of 125

Next, we proceed to compute the second integral. One can use similar trick as in
Example 15 (p. 27) of Chapter 1 to compute this integral.
1
For the 2nd integral ∫ 2 𝑑𝑥, note that
𝑥 −6𝑥+10

1 1 −1
∫ 2 𝑑𝑥 = ∫ 𝑑𝑥 = tan (𝑥 − 3) + 𝐶′
𝑥 − 6𝑥 + 10 (𝑥 − 3)2 + 1

Note:
The last equality follows from the fact that
1 −1
1 1 −1 (
∫ 𝑑𝑥 = tan 𝑥 + 𝐶 ⇒ ∫ 𝑑𝑥 = tan 𝑎𝑥 + 𝑏) + 𝐶.
1 + 𝑥2 1 + (𝑎𝑥 + 𝑏)2 𝑎

Summing up, we have

2𝑥 − 5
∫ 2 𝑑𝑥 = ln|𝑥 2 − 6𝑥 + 10| + tan−1 (𝑥 − 3) + 𝐶.
𝑥 − 6𝑥 + 10
 Page 18 of 125

Example 8
Compute the integral
3𝑥 + 11
∫ 2
𝑑𝑥
2𝑥 + 8𝑥 + 13
Solution:
𝑑𝑦
We try the substitution 𝑦 = 2𝑥 2 + 8𝑥 + 13. Then we have = 4𝑥 + 8.
𝑑𝑥

Following the idea of Example 7, the integral is then rewritten as

3𝑥+11
=(3𝑥+6)+5
3𝑥 + 11 3𝑥 + 6 5
∫ 2 𝑑𝑥 =
⏞ ∫ 2 𝑑𝑥 + ∫ 2 𝑑𝑥
2𝑥 + 8𝑥 + 13 2𝑥 + 8𝑥 + 13 2𝑥 + 8𝑥 + 13
substitute 𝑦=2𝑥 2 +8𝑥+13
⏞ 3𝑥 + 6 1 5
=∫ ( ) 𝑑𝑦 + ∫ 2 𝑑𝑥
2𝑥 2 + 8𝑥 + 13 4𝑥 + 8 2𝑥 + 8𝑥 + 13
3 1 1
= ∫ 𝑑𝑦 + 5 ∫ 𝑑𝑥
4 𝑦 ( )
2 𝑥+2 +52
 Page 19 of 125

3 5 1
| |
= ln 𝑦 + ∫ 𝑑𝑥
4 5 2 (𝑥 + 2)2 + 1
5
2
2 2
(𝑥+2)2 =[√ (𝑥+2)]
5 5
3 1
=
⏞ ln|2𝑥 2 + 8𝑥 + 13| + ∫ 2 𝑑𝑥
4 2 2√2
[(√ 𝑥 + )] + 1
5 √5

3 1 2 2√2
= ln|2𝑥 2 + 8𝑥 + 13| + tan−1 (√ 𝑥 + ) +𝐶
4 5 √5
√2
( 5 )
3 5 2 2√2
2
= ln|2𝑥 + 8𝑥 + 13| + √ −1 √
tan ( 𝑥 + ) + 𝐶.
4 2 5 √5
 Page 20 of 125

Example 9 (Very Tricky, you just need to know the result)

Compute the integral

∫ sec 𝑥 𝑑𝑥.

Solution:
To compute this integral, we multiply both numerator and denominator by a factor of
(sec 𝑥 + tan 𝑥 ), i.e.,
sec 𝑥 (sec 𝑥 + tan 𝑥 ) sec 2 𝑥 + sec 𝑥 tan 𝑥
∫ sec 𝑥 𝑑𝑥 = ∫ 𝑑𝑥 = ∫ 𝑑𝑥.
sec 𝑥 + tan 𝑥 sec 𝑥 + tan 𝑥
𝑑𝑦
We then let 𝑦 = sec 𝑥 + tan 𝑥. Then = sec 𝑥 tan 𝑥 + sec 2 𝑥 and the integral
𝑑𝑥
becomes
sec 2 𝑥 + sec 𝑥 tan 𝑥 sec 2 𝑥 + sec 𝑥 tan 𝑥 1
∫ 𝑑𝑥 = ∫ ( 2
) 𝑑𝑦
sec 𝑥 + tan 𝑥 sec 𝑥 + tan 𝑥 sec 𝑥 tan 𝑥 + sec 𝑥
 Page 21 of 125

1
= ∫ 𝑑𝑦 = ln |𝑦| + 𝐶
𝑦
= ln|sec 𝑥 + tan 𝑥 | + 𝐶.

Remark of Example 9
Using similar trick, one can show that

∫ csc 𝑥 𝑑𝑥 = − ln|csc 𝑥 + cot 𝑥 | + 𝐶.

(Hint: Multiply both numerator and denominator of the integrand by (csc 𝑥 + cot 𝑥).)
 Page 22 of 125

Computing integral involving trigonometric functions: ∫ sin𝑚 𝜃 cos n 𝜃 𝑑𝜃

Example 10 (𝒎 = 𝐨𝐝𝐝) Compute the integral
∫ sin5 𝑥 𝑑𝑥.

IDEA:
Although one can compute this integral by using product-to-sum formula, it
would be quite tedious since we need to use this formula for 4 times.
Solution:
𝑑𝑦
We let 𝑦 = cos 𝑥. Then = − sin 𝑥 and the integral becomes
𝑑𝑥
5 5
1
∫ sin 𝑥 𝑑𝑥 = ∫ sin 𝑥 (− 𝑑𝑦) = − ∫ sin4 𝑥 𝑑𝑦 = − ∫(1 − cos2 𝑥)2 𝑑𝑦
sin 𝑥
2 2 2 4
2𝑦 3 𝑦 5
= − ∫(1 − 𝑦 ) 𝑑𝑦 = − ∫(1 − 2𝑦 + 𝑦 )𝑑𝑦 = −𝑦 + − +𝐶
3 5
2 cos 3 𝑥 cos 5 𝑥
= − cos 𝑥 + − + 𝐶.
3 5
 Page 23 of 125

Example 11 (𝒏 = 𝐨𝐝𝐝) Compute the integral

∫ cos 3 𝑥 sin2 𝑥 𝑑𝑥

Solution:
𝑑𝑦
Let 𝑦 = sin 𝑥, then = cos 𝑥 and the integral becomes
𝑑𝑥

1
∫ cos 3 𝑥 sin2 𝑥 𝑑𝑥 = ∫ cos 3 𝑥 sin2 𝑥 ( 𝑑𝑦) = ∫ cos2 𝑥 sin2 𝑥 𝑑𝑦
cos 𝑥
= ∫(1 − sin2 𝑥 ) sin2 𝑥 𝑑𝑦 = ∫ 𝑦 2 (1 − 𝑦 2 )𝑑𝑦 = ∫(𝑦 2 − 𝑦 4 )𝑑𝑦
𝑦3 𝑦5 sin3 𝑥 sin5 𝑥
= − +𝐶 = − + 𝐶.
3 5 3 5
Remark: (𝒎, 𝒏 = 𝐞𝐯𝐞𝐧) Use double-angle formulae:
1 − cos 2𝜃 1 + cos 2𝜃
sin2 𝜃 = , cos2 𝜃 = .
2 2
 Page 24 of 125

Trigonometric substitution
Sometimes, using method of substitution with simple substitution may not be
successful. As an example, we consider the following integral
1
∫ 3 𝑑𝑥 , 𝑥>0
(1 + 𝑥 2 )2
As a starting point, one may try to use the substitution 𝑦 = 1 + 𝑥 2 to simplify the
1
integrand as 3 .
𝑦2

𝑑𝑦 1
Since = 2𝑥 ⇒ 𝑑𝑥 = 𝑑𝑦, then the integral is then transformed into
𝑑𝑥 2𝑥

1 1 1 1
∫ 3 𝑑𝑥 = ∫ 3 𝑑𝑦 = 2 ∫ 3 𝑑𝑦.
(1 + 𝑥 2 )2 2𝑥 (1 + 𝑥 2 )2 𝑦 2 √𝑦 − 1
It appears that the integral is still hard to integrate after transformation.
One has to try other substitution that can simplify the integral significantly.
 Page 25 of 125

Recall that the main purpose of doing substitution is to transform the denominator
(which is the sum of two terms originally) into a single term which is good for
algebraic manipulation.
An alternative approach (perhaps the most popular one) is to transform the variable
𝑥 into some trigonometric functions (say 𝑥 = sin 𝜃 , cos 𝜃 , tan 𝜃 , sec 𝜃 etc.,) so
that the terms in the denominator can be combined using trigonometric identities.
𝑑𝑥
In this example, we take 𝑥 = tan 𝜃 (⇒ = sec 2 𝜃) and the integral becomes
𝑑𝜃

1 1 1+tan2 𝜃=sec2 𝜃 sec 2 𝜃

( 2
∫ 3 𝑑𝑥 = ∫ 3 sec 𝜃 𝑑𝜃) =
⏞ ∫ 𝑑𝜃
sec 3 𝜃
(1 + 𝑥 2 )2 (1 + tan2 𝜃)2
1
sec 𝜃=
1 cos 𝜃
=∫ 𝑑𝜃 =
⏞ ∫ cos 𝜃 𝑑𝜃 = sin 𝜃 + 𝐶 = ⋯.
sec 𝜃
 Page 26 of 125

The following identities may be useful for trigonometric substitution:

1 − cos2 𝜃 = sin2 𝜃
1 − sin2 𝜃 = cos 2 𝜃
1 + tan2 𝜃 = sec 2 𝜃
sec 2 𝜃 − 1 = tan2 𝜃
1 + cot 2 𝜃 = csc 2 𝜃
csc 2 𝜃 − 1 = cot 2 𝜃
 Page 27 of 125

Example 12
Compute the integral

∫ √3 − 𝑥 2 𝑑𝑥 , −√3 < 𝑥 < √3

Note:
Simple substitution such as 𝑦 = 3 − 𝑥 2 does not work in this
𝑑𝑦 1
context. Note that = −2𝑥 ⇒ 𝑑𝑥 = − 𝑑𝑦 and the integral
𝑑𝑥 2𝑥
becomes
1 1 1
∫ √3 − 𝑥 2 𝑑𝑥 = ∫ √3 − 𝑥 2 (− ) 𝑑𝑦 = − ∫ √𝑦𝑑𝑦
2𝑥 2 √3 − 𝑦

Solution:
𝑑𝑥
We take 𝑥 = √3 sin 𝜃 (so that 𝑥 2 = 3 sin2 𝜃), then = √3 cos 𝜃 and the integral
𝑑𝜃
becomes
 Page 28 of 125

∫ √3 − 𝑥 2 𝑑𝑥 = ∫ √3 − 3 sin2 𝜃 (√3 cos 𝜃 𝑑𝜃)

= ∫ 3 cos 2 𝜃 𝑑𝜃 = 3 ∫(cos 𝜃)(cos 𝜃)𝑑𝜃

1
= 3 ∫ [cos(𝜃 + 𝜃) + cos(𝜃 − 𝜃)]𝑑𝜃
2
3 3
= ∫ cos 2𝜃 𝑑𝜃 + ∫ 1𝑑𝜃
2 2
3 1 3 3 3
= ( sin 2𝜃) + 𝜃 + 𝐶 = sin 2𝜃 + 𝜃 + 𝐶
2 2 2 4 2
𝑥=√3 sin 𝜃
𝑥
⇒ 𝜃=sin−1
√3 3 𝑥 3 −1 𝑥
−1
=
⏞ sin (2 sin ) + sin + 𝐶.
4 √3 2 √3
 Page 29 of 125

Remark or Example 12:

Alternatively, one can transform the final answer into the function of 𝑥 in as follows:
sin(𝐴+𝐵)
3 3 =sin 𝐴 cos 𝐵+cos 𝐴 sin 𝐵 3 3
sin 2𝜃 + 𝜃 + 𝐶 =
⏞ 2 sin 𝜃 cos 𝜃 + 𝜃 + 𝐶
4 2 4 2
3 𝑥 √3 − 𝑥 2 3 −1 𝑥
= ( ) + sin +𝐶
2 √3 √3 2 √3
𝑥√3−𝑥 2 3 𝑥
= + sin−1 + 𝐶 (better)
2 2 √3

√3 𝑥

𝜃
√3 − 𝑥 2
 Page 30 of 125

Example 13
Compute the integral
1
∫ 𝑑𝑥
𝑥 3 √𝑥 2 −1
Solution:
𝑑𝑥
Let 𝑥 = sec 𝜃 (so that 𝑥 2 − 1 = sec 2 𝜃 − 1 = tan2 𝜃), then = sec 𝜃 tan 𝜃 and
𝑑𝜃
the integral becomes
1 1
∫ 𝑑𝑥 = ∫ (sec 𝜃 tan 𝜃 𝑑𝜃)
𝑥 3 √𝑥 2 −1 sec 3 𝜃 √sec 2 𝜃 −1
1 1
=∫ ( )
sec 𝜃 tan 𝜃 𝑑𝜃 = ∫ 𝑑𝜃
sec 3 𝜃 tan 𝜃 sec 2 𝜃
2
1
= ∫ cos 𝜃 𝑑𝜃 = ∫ [cos(𝜃 + 𝜃) + cos(𝜃 − 𝜃 )]𝑑𝜃
2
1
= ∫[cos 2𝜃 + 1]𝑑𝜃
2
 Page 31 of 125

1 sin 2𝜃 1 1
= ( + 𝜃) + 𝐶 = sin 2𝜃 + 𝜃 + 𝐶
2 2 4 2
1 1
= (2 sin 𝜃 cos 𝜃) + 𝜃 + 𝐶
4 2
1 √𝑥 2 − 1 1 1
= ( ) ( ) + sec −1 𝑥 + 𝐶
2 𝑥 𝑥 2

√𝑥 2 − 1 1 −1
= 2
+ sec 𝑥 + 𝐶.
2𝑥 2
 Page 32 of 125

Example 14 (A bit tricky)

Compute the integral
1
∫ 𝑑𝑥
√𝑥 2 + 4𝑥 + 7
Note:

One has to “combine” the terms 𝑥 2 and 4𝑥 into a single square using
completing square technique.

Solution
Using the completing square technique, we have
1 1 1
∫ 𝑑𝑥 = ∫ 𝑑𝑥 = ∫ 𝑑𝑥
√𝑥 2 + 4𝑥 + 7 √(𝑥 2 + 4𝑥 + 4) + 3 √(𝑥 + 2)2 +3
 Page 33 of 125

We then adopt the trigonometric substitution and let 𝑥 + 2 = √3 tan 𝜃 (so that
(𝑥 + 2)2 + 3 = 3 tan2 𝜃 + 3 = 3(tan2 𝜃 + 1) = 3 sec 2 𝜃).
𝑑𝑥
Then 𝑥 = √3 tan 𝜃 − 2 and = √3 sec 2 𝜃. The integral is transformed into
𝑑𝜃
1 1
∫ 𝑑𝑥 = ∫ (√3 sec 2 𝜃 𝑑𝜃)
√(𝑥 + 2)2 + 3 √3 sec 2 𝜃

= ∫ sec 𝜃 𝑑𝜃 = ln|sec 𝜃 + tan 𝜃| + 𝐶

= ln |√1 + tan2 𝜃 + tan 𝜃| + 𝐶

2
𝑥+2 𝑥+2
= ln |√1 + ( ) + |+𝐶
√3 √3

𝑥 2 + 4𝑥 + 7 𝑥 + 2
= ln |√ + | + 𝐶 = ln |√𝑥 2 + 4𝑥 + 7 + 𝑥 + 2| + 𝐶 ′ .
3 √3
 Page 34 of 125

How to choose the suitable substitution?

So far, we have two different types of substitution (standard one and the
trigonometric one). One may ask a question that which substitution we should take
when using the method substitution. In fact, there is no clear guideline on the choice
the substitution. Here’s my suggestion for beginners:
 It is always better to start with a standard, straight-forward substitution (see
Example 1 – 6) such that the computation process is easier.

 In the unlucky case when the standard substitution does not work. One can try
to use some algebraic trick (see Example 7 – 8) to transform the integral or try
to adopt trigonometric substitution (see Example 12-14).
 Page 35 of 125

As an illustration, we consider the following two integrals

1 𝑥5
∫ 𝑑𝑥 , ∫ 2 𝑑𝑥.
(1 + 𝑥 2 )2 (𝑥 − 1)3

Computation of the first integral

For the first integral, one can first try the substitution
𝑦 = 1 + 𝑥 2.
𝑑𝑦 1
Then = 2𝑥 ⇒ 𝑑𝑥 = 𝑑𝑦 and the integral becomes
𝑑𝑥 2𝑥

1 1 1 1 1
∫ 𝑑𝑥 = ∫ ( 𝑑𝑦) = ∫ 𝑑𝑦
(1 + 𝑥 2 )2 (1 + 𝑥 2 )2 2𝑥 2 𝑦 2 √𝑦 − 1

which is difficult to integrate.

 Page 36 of 125

In this case, one can try other substitutions. Here, we use trigonometric substitution
and let 𝑥 = tan 𝜃 (so that 1 + 𝑥 2 = 1 + tan2 𝜃 = sec 2 𝜃).
𝑑𝑥
Then = sec 2 𝜃 and the integral becomes
𝑑𝜃

1 1 2
1 2
∫ 𝑑𝑥 = ∫ (sec 𝜃 𝑑𝜃 ) = ∫ 𝑑𝜃 = ∫ cos 𝜃 𝑑𝜃
(1 + 𝑥 2 )2 (1 + tan2 𝜃)2 sec 2 𝜃
1 1
= ∫[cos(𝜃 + 𝜃) + cos(𝜃 − 𝜃)]𝑑𝜃 = ∫(cos 2𝜃 + 1)𝑑𝜃
2 2
1 𝜃 𝑥 tan−1 𝑥
= sin 2𝜃 + + 𝐶 = ⋯ = + + 𝐶.
4 2 2(1 + 𝑥 2 ) 2
Computation of the second integral
For the second integral, we may try the substitution
𝑦 = 𝑥 2 − 1.
𝑑𝑦 1
Then = 2𝑥 ⇒ 𝑑𝑥 = 𝑑𝑦 and the integral becomes
𝑑𝑥 2𝑥
 Page 37 of 125

𝑥5 𝑥5 1 1 𝑥4
∫ 2 𝑑𝑥 = ∫ 2 ( 𝑑𝑦) = ∫ 2 𝑑𝑦
(𝑥 − 1)3 (𝑥 − 1)3 2𝑥 2 ( 𝑥 −1 )3

𝑦=𝑥 2 −1
⇒ 𝑥 2 =𝑦+1
1 (𝑦 + 1)2 1 𝑦 2 + 2𝑦 + 1
=
⏞ ∫ 𝑑𝑦 = ∫ 𝑑𝑦
2 𝑦3 2 𝑦3
1 1 1
= ∫ 𝑑𝑦 + ∫ 𝑦 −2 𝑑𝑦 + ∫ 𝑦 −3 𝑑𝑦
2 𝑦 2
1 𝑦 −2+1 1 𝑦 −3+1
= ln|𝑦| + + ( )+𝐶
2 −2 + 1 2 −3 + 1
1 2
1 1
= ln|𝑥 − 1| − 2 − +𝐶
2 𝑥 − 1 4(𝑥 2 − 1)2
Since the simple substitution is OK for this situation, there is no need to use
trigonometric substitution which may complicate the matter.
 Page 38 of 125

Integration by parts
This method is first motivated from the product rule in differentiation: Let 𝑢 = 𝑢(𝑥)
and 𝑣 = 𝑣(𝑥) be two differentiable functions in 𝑥, then
𝑑 𝑑𝑢 𝑑𝑣
𝑢𝑣 = 𝑣 +𝑢 … … (∗).
𝑑𝑥 𝑑𝑥 𝑑𝑥
We integrate the equation (∗) with respect to 𝑥, we get
𝑑(𝑢𝑣) 𝑑𝑢 𝑑𝑣
∫ 𝑑𝑥 = ∫ 𝑣 𝑑𝑥 + ∫ 𝑢 𝑑𝑥
𝑑𝑥 𝑑𝑥 𝑑𝑥

⇒ ∫ 𝑑(𝑢𝑣) = ∫ 𝑣𝑑𝑢 + ∫ 𝑢𝑑𝑣

⇒ 𝑢𝑣 = ∫ 𝑣𝑑𝑢 + ∫ 𝑢𝑑𝑣
Recall that ∫ 1𝑑𝑦 = 𝑦 + 𝐶
and we take 𝑦 = 𝑢𝑣
⇒ ∫ 𝑢𝑑𝑣 = 𝑢𝑣 − ∫ 𝑣𝑑𝑢.
 Page 39 of 125

One can use the similar integral to obtain the integration by parts formula for
definite integral. We integrate the equation (∗) both sides with respect to 𝑥 (from
𝑎 to 𝑏):
𝑏 𝑏 𝑏
𝑑(𝑢𝑣) 𝑑𝑢 𝑑𝑣
∫ 𝑑𝑥 = ∫ 𝑣 𝑑𝑥 + ∫ 𝑢 𝑑𝑥
𝑎 𝑑𝑥 𝑎 𝑑𝑥 𝑎 𝑑𝑥
𝑏 𝑏 𝑏
⇒ ∫ 𝑑 (𝑢𝑣 ) = ∫ 𝑣𝑑𝑢 + ∫ 𝑢𝑑𝑣
𝑎 𝑎 𝑎
𝑏 𝑏
𝑢(𝑏)𝑣 (𝑏) − 𝑢(𝑎)𝑣(𝑎) = ∫ 𝑣𝑑𝑢 + ∫ 𝑢𝑑𝑣
⇒⏟
𝑎 𝑎
=𝑢𝑣|𝑏 𝑏
𝑎 or [𝑢𝑣]𝑎

𝑏 𝑏
⇒ ∫ 𝑢𝑑𝑣 = 𝑢𝑣|𝑏𝑎 − ∫ 𝑣𝑑𝑢.
𝑎 𝑎

The two formulae in the boxes are called integration by parts formulae.
 Page 40 of 125

How do the formulae work?

In order to appreciate this formula, we consider the following integral:

∫ ln 𝑥 𝑑𝑥

Using the integration by parts formula with 𝑢 = ln 𝑥 and 𝑑𝑣 = 𝑑𝑥

(or 𝑣 = ∫ 𝑑𝑥 = 𝑥), we have
𝑑 1 1
∫ ln
⏟ 𝑥 𝑑𝑥 𝑥 𝑑 (⏟ln 𝑥 )
⏟ = 𝑥 ln 𝑥 − ∫ ⏟ 𝑑𝑥
ln 𝑥 = ⇒ 𝑑(ln 𝑥 ) = 𝑑𝑥
𝑥 𝑥
𝑢 𝑑𝑣 𝑣 𝑢

1
= 𝑥 ln 𝑥 − ∫ 𝑥 ( 𝑑𝑥)
𝑥

= 𝑥 ln 𝑥 − ∫ 1𝑑𝑥

= 𝑥 ln 𝑥 − 𝑥 + 𝐶.
 Page 41 of 125

Consider another integral:

∫ 𝑥 sin 𝑥 𝑑𝑥.

Using integration by parts with 𝑢 = 𝑥 and 𝑑𝑣 = sin 𝑥 𝑑𝑥

(so that 𝑣 = ∫ sin 𝑥 𝑑𝑥 = − cos 𝑥), we have

∫⏟
𝑥⏟ 𝑥 𝑑 (⏟− cos 𝑥 ) = ⏟
sin 𝑥 𝑑𝑥 = ∫ ⏟ 𝑥(− cos 𝑥 ) − ∫ (⏟− cos 𝑥 ) 𝑑 ⏟
𝑥
𝑢 𝑑𝑣 𝑢 𝑣 𝑢𝑣 𝑣 𝑢

= −𝑥 cos 𝑥 + ∫ cos 𝑥 𝑑𝑥

= −𝑥 cos 𝑥 + sin 𝑥 + 𝐶.
 Page 42 of 125

Insights about integration by parts

 Similar to the method of substitution, integration by parts aims to transform a
given integral into another integral (with some extra terms) in order to simplify
the calculation:

∫ 𝑥 sin 𝑥 𝑑𝑥 = −𝑥 cos 𝑥 + ∫ cos 𝑥 𝑑𝑥

⏟ ⏟
hard to compute easier

 The principle of this method is to “eliminate” some ugly terms (say ln 𝑥 , 𝑥, … ,

etc.,) through differentiation (since they will be put in the differential operator
after the transform).
1
∫ ln
⏟ 𝑥𝑑⏟ 𝑥 = 𝑥 ln 𝑥 − ∫ ⏟ ( )
𝑥 𝑑 ⏟ln 𝑥 = 𝑥 ln 𝑥 − ∫ 𝑥 ( 𝑑𝑥)
𝑢 𝑣 𝑣 𝑢
𝑥

ln 𝑥 is put inside the

ln 𝑥 is being “eliminated”
operator 𝑑(? ? ).
after differentiation
 Page 43 of 125

How to use integration by parts properly?

The most crucial part of using integration by parts is how to allocate the functions
into 𝑢(𝑥) and 𝑑𝑣. The following guideline (according to my past experience)
suggests one of the ways:
Step 1: Pick 𝑢(𝑥)
 Recall that the function 𝑢(𝑥) will be put inside the differential operator and is
then differentiated after the use of the formula.
𝑑𝑢
∫ 𝑢𝑑𝑣 = 𝑢𝑣 − ∫ 𝑣𝑑𝑢 = 𝑢𝑣 − ∫ 𝑣 ( ) 𝑑𝑥
𝑑𝑥

 Since we wish to simplify the integral as much as possible, we should choose

𝑢(𝑥) in such a way that 𝑢(𝑥) can be either “eliminated completely” or
“transformed into convenient form” through differentiation during the transform.
 Page 44 of 125

The following shows the priority about the choice of 𝑢(𝑥)

Highest Priority 𝑑𝑓1 (𝑥) 1 −1 1
𝑓1 (𝑥 ) = sin −1
𝑥 , cos −1
𝑥 or tan −1
𝑥 ⇒ 𝑑𝑥
= ⏟ 2 , ⏟ 2 or
√1−𝑥 √1−𝑥 ⏟2
1+𝑥
sin−1 𝑥 cos−1 𝑥 tan−1 𝑥

𝑑𝑓2 (𝑥 ) 1
𝑓2 (𝑥 ) = ln 𝑥 ⇒ =
𝑑𝑥 𝑥
𝑑𝑓3 (𝑥 )
𝑓3 (𝑥 ) = 𝑥 𝑛 , 𝑛 = 1,2,3,4, … ⇒ = 𝑛𝑥 𝑛−1
𝑑𝑥
(𝑛 must be positive integer!!)
𝑑𝑓4 (𝑥 )
𝑓4 (𝑥 ) = sin 𝑥 , cos 𝑥 ⇒ = cos
⏟𝑥 or ⏟− sin 𝑥
𝑑𝑥 sin 𝑥 cos 𝑥
Lowest Priority
Step 2: Find 𝑣(𝑥)
After 𝑢(𝑥) is identified, the remaining part in the integrand (say 𝑓(𝑥 )𝑑𝑥) will be
𝑑𝑣. The function 𝑣(𝑥) can be computed by
𝑑𝑣 = 𝑓(𝑥 )𝑑𝑥 ⇒ ∫ 𝑑𝑣 = ∫ 𝑓(𝑥 )𝑑𝑥 ⇒ 𝑣 = ∫ 𝑓(𝑥 )𝑑𝑥.
 Page 45 of 125

Example 15
Compute the integral

∫ 𝑥 4 ln 𝑥 𝑑𝑥.

IDEA:
One has to eliminate one of the functions 𝑥 4 and ln 𝑥 using integration
by parts. Although both 𝑥 4 and ln 𝑥 can be “eliminated” through
differentiation, we shall choose 𝑢(𝑥 ) = ln 𝑥 since it has a higher priority.
(It is because we are able to integrate 𝑥 4 )

Solution:
4 4 𝑥5
We pick 𝑢 = 𝑢(𝑥 ) = ln 𝑥. Then 𝑑𝑣 = 𝑥 𝑑𝑥 ⇒ 𝑣 = ∫ 𝑥 𝑑𝑥 = .
5
 Page 46 of 125

Using integration by parts, we have

5
𝑥
∫ 𝑥 4 ln 𝑥 𝑑𝑥 = ∫ ln
⏟ (𝑥 4 𝑑𝑥 ) = ∫ ln 𝑥 𝑑 ( )
𝑥⏟
𝑢 𝑑𝑣 ⏟ 5
∫ 𝑢𝑑𝑣

𝑥5 𝑥5
= ln 𝑥 − ∫ 𝑑 (ln 𝑥 )
⏟
5 ⏟5
𝑢𝑣 ∫ 𝑣𝑑𝑢

𝑥5 𝑥5 1
= ln 𝑥 − ∫ ( 𝑑𝑥)
5 5 𝑥
𝑥5 1 4
𝑥5 𝑥5
= ln 𝑥 − ∫ 𝑥 𝑑𝑥 = ln 𝑥 − + 𝐶.
5 5 5 25
 Page 47 of 125

Example 16
Compute the integral

∫ 𝑥 tan−1 𝑥 𝑑𝑥.

IDEA:
Since there is an inverse trigonometric function in the integrand, one has
to eliminate tan−1 𝑥 first.
Solution:
−1 𝑥2
We pick 𝑢 = 𝑢(𝑥 ) = tan 𝑥 and 𝑑𝑣 = 𝑥𝑑𝑥 ⇒ 𝑣 = ∫ 𝑥𝑑𝑥 = .
2
From integration by parts, we have
2
𝑥
∫ 𝑥 tan−1 𝑥 𝑑𝑥 = ∫ ⏟
tan−1 𝑥 (⏟𝑥𝑑𝑥 ) = ∫ tan−1 𝑥 𝑑 ( )
𝑢 𝑑𝑣 ⏟ 2
∫ 𝑢𝑑𝑣
 Page 48 of 125

𝑥2 𝑥2
= tan 𝑥 − ∫ 𝑑(tan−1 𝑥 )
−1
⏟
2 ⏟2
𝑢𝑣 ∫ 𝑣𝑑𝑢

𝑥2 −1
𝑥2 1 𝑥2 −1
1 𝑥2
= tan 𝑥 − ∫ ( 2
) 𝑑𝑥 = tan 𝑥 − ∫ 2
𝑑𝑥
2 2 1+𝑥 2 2 1+𝑥

(∗)
𝑥2 −1
1 𝑥2 + 1 1
= tan 𝑥 − ∫ [ 2 − 2 ] 𝑑𝑥
2 2 𝑥 +1 𝑥 +1
𝑥2 −1
1 1 1
= tan 𝑥 − ∫ 1𝑑𝑥 + ∫ 2 𝑑𝑥
2 2 2 𝑥 +1
𝑥2 𝑥 1
= tan 𝑥 − + tan−1 𝑥 + 𝐶.
−1
2 2 2
Remark of (∗)
The purpose of doing such decomposition is to eliminate the 𝑥 2 in the
numerator.
Example 17
 Page 49 of 125

Compute

∫ 𝑥 2 𝑒 3𝑥 𝑑𝑥

Solution:
1
Pick 𝑢 = 𝑥 2 and 𝑑𝑣 = 𝑒 3𝑥 𝑑𝑥 ⇒ 𝑣 = ∫ 𝑒 3𝑥 𝑑𝑥 = 𝑒 3𝑥 .
3

From integration by parts, we have

3𝑥
𝑒
∫ 𝑥 2 𝑒 3𝑥 𝑑𝑥 = ∫ 𝑥
⏟2 (⏟𝑒 3𝑥 𝑑𝑥 ) = ∫ 𝑥 2 𝑑 ( )
𝑢 𝑑𝑣 ⏟ 3
∫ 𝑢𝑑𝑣

𝑥 2 𝑒 3𝑥 𝑒 3𝑥
= −∫ 𝑑 (𝑥 2 )
⏟3 ⏟ 3
𝑢𝑣 ∫ 𝑣𝑑𝑢

𝑥 2 𝑒 3𝑥 𝑒 3𝑥 𝑥 2 𝑒 3𝑥 2
= −∫ (2𝑥𝑑𝑥 ) = − ∫ 𝑥𝑒 3𝑥 𝑑𝑥
3 3 3 3
 Page 50 of 125

To continue the computation, we apply integration by parts again on the 2nd integral.
1
Take 𝑢 = 𝑥 and 𝑑𝑣 = 𝑒 3𝑥 𝑑𝑥 ⇒ 𝑣 = ∫ 𝑒 3𝑥 𝑑𝑥 = 𝑒 3𝑥 .
3
2 3𝑥
2 3𝑥
𝑥 𝑒 2
∫ 𝑥 𝑒 𝑑𝑥 = − ∫ 𝑥𝑒 3𝑥 𝑑𝑥
3 3

𝑥 2 𝑒 3𝑥 2 𝑒 3𝑥 𝑥 2 𝑒 3𝑥 2 𝑥𝑒 3𝑥 𝑒 3𝑥
= − ∫𝑥 𝑑( ) = − [ −∫ 𝑑𝑥 ]
3 3⏟ 3 3 3 3⏟ ⏟ 3
𝑢𝑣 ∫ 𝑣𝑑𝑢
∫ 𝑢𝑑𝑣

𝑥 2 𝑒 3𝑥 2 3𝑥 2
= − 𝑥𝑒 + ∫ 𝑒 3𝑥 𝑑𝑥
3 9 9
𝑥 2 𝑒 3𝑥 2𝑥𝑒 3𝑥 2
= − + 𝑒 3𝑥 + 𝐶.
3 9 27
 Page 51 of 125

Example 18 (Integration by parts for definite integral)

Compute the integral
𝜋
2
∫ 𝑥 sin 𝑥 𝑑𝑥
0

Solution:
We take 𝑢(𝑥 ) = 𝑥. Then 𝑑𝑣 = sin 𝑥 𝑑𝑥 ⇒ 𝑣 = ∫ sin 𝑥 𝑑𝑥 = − cos 𝑥.
From integration by parts, we have
𝜋 𝜋 𝜋
2 2 2
∫ 𝑥 sin 𝑥 𝑑𝑥 = ∫ ⏟𝑥⏟(sin 𝑥 𝑑𝑥 ) = ∫ 𝑥𝑑 (− cos 𝑥 )
0 0 𝑢 𝑑𝑣 ⏟0
∫ 𝑢𝑑𝑣
𝜋
𝜋 𝜋
2
= [⏟−𝑥 cos 𝑥 ]02 − ∫ (− cos 𝑥 )𝑑𝑥 = 0 + [⏟sin 𝑥 ]02 = 1.
𝜋 𝜋 0 𝜋
=− cos −(−0 cos 0) sin −sin 0
2 2 2
 Page 52 of 125

Harder Examples
In this section, we will present some advanced (not straight forward) applications of
integration by parts.
Difficult situation #1: The function in the integrand cannot be “eliminated” by
differentiation.
Example 19
Compute the integral
∫ 𝑒 3𝑥 sin 𝑥 𝑑𝑥

IDEA:
Unfortunately, both 𝑒 3𝑥 and sin 𝑥 cannot be eliminated through
differentiation. However, sin 𝑥 can be “transformed” into cos 𝑥 after
differentiation. Hence, we still use integration by parts to transform the
integral into another form.
 Page 53 of 125

Solution:
We try to transform the integral by using integration by parts. Take 𝑢 = sin 𝑥 and
1
𝑑𝑣 = 𝑒 3𝑥 𝑑𝑥 ⇒ 𝑣 = ∫ 𝑒 3𝑥 𝑑𝑥 = 𝑒 3𝑥 , then we have
3
1 3𝑥
∫ 𝑒 3𝑥 sin 𝑥 𝑑𝑥 = ∫ sin
⏟𝑥 (𝑒
⏟ 3𝑥
𝑑𝑥) = ∫ sin 𝑥 𝑑 ( 𝑒 )
𝑢
⏟ 3
𝑑𝑣
∫ 𝑢𝑑𝑣
1 3𝑥 1 3𝑥
= 𝑒 sin 𝑥 − ∫ 𝑒 𝑑 (sin 𝑥 )
3 3
1 1
= 𝑒 3𝑥 sin 𝑥 − ∫ 𝑒 3𝑥 cos 𝑥 𝑑𝑥
3 3
To continue, we use the integration by parts again by taking 𝑢 = cos 𝑥 and 𝑑𝑣 =
1
𝑒 3𝑥 𝑑𝑥 ⇒ 𝑣 = ∫ 𝑒 3𝑥 𝑑𝑥 = 𝑒 3𝑥 , then we have
3
1 3𝑥 1 𝑒 3𝑥
= 𝑒 sin 𝑥 − ∫ cos 𝑥 𝑑 ( )
3 3⏟ 3
∫ 𝑢𝑑𝑣
 Page 54 of 125

1 3𝑥 1 𝑒 3𝑥 𝑒 3𝑥
= 𝑒 sin 𝑥 − [cos 𝑥 ( ) − ∫ 𝑑 (cos 𝑥 )]
3 3 3 3
1 1 1
= 𝑒 3𝑥 sin 𝑥 − 𝑒 3𝑥 cos 𝑥 − ∫ 𝑒 3𝑥 sin 𝑥 𝑑𝑥
3 9 9
Summing up, we have

3𝑥
1 3𝑥 1 3𝑥 1
∫𝑒 sin 𝑥 𝑑𝑥 = 𝑒 sin 𝑥 − 𝑒 cos 𝑥 − ∫ 𝑒 3𝑥 sin 𝑥 𝑑𝑥
3 9 9
10 1 1
⇒ ∫ 𝑒 3𝑥 sin 𝑥 𝑑𝑥 = 𝑒 3𝑥 sin 𝑥 − 𝑒 3𝑥 cos 𝑥
9 3 9

3𝑥
3 3𝑥 1 3𝑥
⇒ ∫𝑒 sin 𝑥 𝑑𝑥 = 𝑒 sin 𝑥 − 𝑒 cos 𝑥 + 𝐶.
10 10
 Page 55 of 125

Example 20
Compute the integral

∫ sec 3 𝑥 𝑑𝑥

IDEA:
Different from the integral ∫ sin3 𝑥 𝑑𝑥 that you have seen in Lecture
Note 1 (Review example 3 p. 49), there is no “product-to-sum” formula
for the trigonometric function sec 𝜃. Again, we shall transform the
integral using integration by parts.

 However, one should NOT use the integration by parts directly as follows:
⏟ 3 𝑥 𝑑𝑥
∫ sec ⏟ = 𝑥 sec 3 𝑥 − ∫ 𝑥𝑑(sec 3 𝑥 ) = 𝑥 sec 3 𝑥 − 3 ∫ ⏟
𝑥 sec 3 𝑥 tan 𝑥 𝑑𝑥
𝑢 𝑑𝑣 extra!
 Page 56 of 125

To avoid this problem, one has to ensure that both 𝑢(𝑥) and 𝑣(𝑥) are
trigonometric functions. Note that ∫ sec 2 𝑥 𝑑𝑥 = tan 𝑥, one can use integration by
parts in the following ways:
Let 𝑢 = sec 𝑥 and 𝑑𝑣 = sec 2 𝑥 𝑑𝑥 ⇒ 𝑣 = ∫ sec 2 𝑥 𝑑𝑥 = tan 𝑥.

∫ sec 3 𝑥 𝑑𝑥 = ∫ sec
⏟𝑥 (⏟sec 2 𝑥 𝑑𝑥 ) = ∫ sec 𝑥 𝑑 (tan 𝑥 )
𝑢 𝑑𝑣
⏟
∫ 𝑢𝑑𝑣

= sec 𝑥 tan 𝑥 − ∫ tan 𝑥 𝑑 (sec 𝑥 )

𝑑
sec 𝑥=sec 𝑥 tan 𝑥
𝑑𝑥
=
⏞ sec 𝑥 tan 𝑥 − ∫ tan 𝑥 (sec 𝑥 tan 𝑥 𝑑𝑥 )

= sec 𝑥 tan 𝑥 − ∫ sec 𝑥 tan2 𝑥 𝑑𝑥

1+tan2 𝑥=sec2 𝑥
=
⏞ sec 𝑥 tan 𝑥 − ∫ sec 𝑥 (sec 2 𝑥 − 1)𝑑𝑥
 Page 57 of 125

= sec 𝑥 tan 𝑥 + ∫ sec 𝑥 𝑑𝑥 − ∫ sec 3 𝑥 𝑑𝑥

∫ sec 𝑥𝑑𝑥
=ln|sec 𝑥+tan 𝑥|+𝐶
=
⏞ sec 𝑥 tan 𝑥 + ln|sec 𝑥 + tan 𝑥 | − ∫ sec 3 𝑥 𝑑𝑥

Summing up, we get

∫ sec 3 𝑥 𝑑𝑥 = sec 𝑥 tan 𝑥 + ln|sec 𝑥 + tan 𝑥 | − ∫ sec 3 𝑥 𝑑𝑥

⇒ 2 ∫ sec 3 𝑥 𝑑𝑥 = sec 𝑥 tan 𝑥 + ln|sec 𝑥 + tan 𝑥 |

1
⇒ ∫ sec 3 𝑥 𝑑𝑥 = (sec 𝑥 tan 𝑥 + ln|sec 𝑥 + tan 𝑥 |) + 𝐶.
2
 Don’t forget to add an arbitrary constant 𝐶 to an indefinite integral.
 Page 58 of 125

Difficult situation #2: The integrand involved the complicated function.

Example 21
Compute the integral

∫ 𝑥 ln(2√𝑥 + 1)𝑑𝑥

IDEA:
Due to the existence of logarithmic function ln 𝑥, one may prefer to use
integration by parts with 𝑢 = ln(2√𝑥 + 1). However, the process is tedious.
𝑥2
⏟ √𝑥 + 1) (⏟𝑥𝑑𝑥 ) = ∫ ln(2√𝑥 + 1) 𝑑 ( 2 )
∫ ln(2
𝑢 𝑑𝑣
𝑥2 𝑥2
= ln(2√𝑥 + 1) − ∫ 𝑑(ln(2√𝑥 + 1))
2 2
𝑥2 𝑥2 1
= ln(2√𝑥 + 1) − ∫ ( ) 𝑑𝑥
2 2 √𝑥(2√𝑥 + 1)
 Page 59 of 125

Solution:
One has to use the method of substitution first to simplify the integral. (Transform
the function ln(2√𝑥 + 1) into ln 𝑦.
1 1
𝑑𝑦 𝑑 −
We let 𝑦 = 2√𝑥 + 1. Then = (2𝑥 + 1) = 𝑥
2 2 and the integral becomes
𝑑𝑥 𝑑𝑥

1 3
∫ 𝑥 ln(2√𝑥 + 1)𝑑𝑥 = ∫ 𝑥 ln(2√𝑥 + 1) ( 1 𝑑𝑦) = ∫ 𝑥 2 ln(2√𝑥 + 1) 𝑑𝑦
𝑥 −2
𝑦−1 3
= ∫( ) ln 𝑦 𝑑𝑦
2
1 3
3 2
3 1
= ∫ 𝑦 ln 𝑦 𝑑𝑦 − ∫ 𝑦 ln 𝑦 𝑑𝑦 + ∫ 𝑦 ln 𝑦 𝑑𝑦 − ∫ ln 𝑦 𝑑𝑦
8 8 8 8
Each of these four integrals can be computed using the technique used in Example
15. (Left as exercise.) For your reference, the final answer is given by
 Page 60 of 125

∫ 𝑥 ln(2√𝑥 + 1)𝑑𝑥

1 𝑦4 𝑦4 3 𝑦3 𝑦3 3 𝑦2 𝑦2 1
= [ ln 𝑦 − ] − [ ln 𝑦 − ] + [ ln 𝑦 − ] − [𝑦 ln 𝑦 − 𝑦] + 𝐶
8 4 16 8 3 9 8 2 4 8

(*Remember to put 𝑦 = 2√𝑥 + 1. I skip this step to save space.)

Example 22
Compute the integral
1
2 𝑥
𝑒
∫ 𝑑𝑥 .
1 𝑥3
IDEA:
Similar to Example 22, one needs to simplify the integral using the method of
1
substitution to transform the function 𝑒 into “standard” exponent function
𝑥

𝑒𝑦.
 Page 61 of 125

Solution
1 𝑑𝑦 1
We let 𝑦 = . Then =− and the integral becomes
𝑥 𝑑𝑥 𝑥2
1
when 𝑥=2, 𝑦=
1 2 1
when 𝑥=1, 𝑦=1 1 1
2 𝑥
𝑒 2 𝑒 𝑥
2
2 1 1
∫ 𝑑𝑥 =
⏞ ∫ (−𝑥 𝑑𝑦) = − ∫ 𝑒𝑥 ( ) 𝑑𝑦
1 𝑥3 1 𝑥3 1 𝑥
𝑑𝑣=𝑒 𝑦 𝑑𝑦
𝑏 𝑎
∫𝑎 𝑓(𝑥)𝑑𝑥=− ∫𝑏 𝑓(𝑥)𝑑𝑥 1 ⇒ 𝑣=∫ 𝑒 𝑦 𝑑𝑦=𝑒 𝑦 1
=
⏞ ∫ ⏟ 𝑒 𝑦 𝑑𝑦
𝑦⏟ =
⏞ ∫ 𝑦𝑑 (𝑒 𝑦 )
1 1
2 𝑢 𝑑𝑣 2
1
= [⏟𝑦𝑒 𝑦 ]11/2 − ∫ 𝑒 𝑦 𝑑𝑦
1
1 12 2
1
1𝑒 − 𝑒
2

1 1 𝑦 1 1 1 1 1 1
= 𝑒 − 𝑒 2 − [𝑒 ]1 = 𝑒 − 𝑒 2 − (𝑒 − 𝑒 2 ) = 𝑒 2 .
2 2 2 2
 Page 62 of 125

Method of substitution v.s. integration by parts

Summing up, the method of substitution and integration by parts aim to transform
the integral in two different ways:
Method of substitution
1
Transform the functions (say 𝑒 , sin √𝑥 , cos(ln 𝑥 ), etc.,) into “standard functions”
𝑥
1
(say 𝑒 𝑦 , sin 𝑦, cos 𝑦) using the substitution (say 𝑦 = , √𝑥, ln 𝑥)
𝑥

Integration by parts
Transform the functions (say sin−1 𝑥, ln 𝑥) into simple function using differentiation
𝑑𝑢 1 1
(take 𝑢(𝑥 ) = sin−1 𝑥, ln 𝑥 and = 2
, ).
𝑑𝑥 √1−𝑥 𝑥

 This method may not work well for the functions such as ln(2√𝑥 + 1) etc. One
needs to transform these functions into standard one ln 𝑦 using the method of
substitution. (See Example 21 and 22)
 Page 63 of 125

Example 23
Compute the integral
∫ 𝑥 5 cos(3 − 𝑥 3 ) 𝑑𝑥

Solution:
We first simplify the integral using the method of substitution again.
𝑑𝑦
We let 𝑦 = 3 − 𝑥 3 . Then = −3𝑥 2 and the integral becomes
𝑑𝑥

5 3) 5 3)
1
∫ 𝑥 cos(3 − 𝑥 𝑑𝑥 = ∫ 𝑥 cos(3 − 𝑥 ( 𝑑𝑦)
−3𝑥 2
𝑥 3 =3−𝑦
1 1
=
⏞ ( )
− ∫ 3 − 𝑦 cos 𝑦 𝑑𝑦 = − ∫ ⏟
(3 − 𝑦) ⏟
(cos 𝑦 𝑑𝑦)
3 3
𝑢 𝑑𝑣
𝑑𝑣=cos 𝑦𝑑𝑦
⇒ 𝑣=∫ cos 𝑦𝑑𝑦=sin 𝑦
1
=
⏞ − ∫(3 − 𝑦)𝑑 (sin 𝑦)
3
 Page 64 of 125

1
= − [(3 − 𝑦) sin 𝑦 − ∫ sin 𝑦 𝑑 (3 − 𝑦)]
3
𝑑
(3−𝑦)=−1
𝑑𝑦
1
=
⏞ − [(3 − 𝑦) sin 𝑦 − ∫ sin 𝑦 (−𝑑𝑦)]
3
1 1
= − (3 − 𝑦) sin 𝑦 − ∫ sin 𝑦 𝑑𝑦
3 3
1 1
= − (3 − 𝑦) sin 𝑦 + cos 𝑦 + 𝐶
3 3
1 3 1
= − 𝑥 sin(3 − 𝑥 ) + cos(3 − 𝑥 3 ) + 𝐶.
3
3 3
 Page 65 of 125
There is sin−1 𝑥.
Eliminate it using
Example 24
integration by parts first!
Compute the integral
𝑥 sin−1 𝑥 There is 1 − 𝑥 2 in the
∫ 𝑑𝑥
√1 − 𝑥2 denominator. Transform it using
the method of substitution first!
Solution
Method 1: Use integration by parts
first
𝑦=1−𝑥 2
𝑥 𝑥
Let 𝑢(𝑥 ) = sin−1 𝑥, then 𝑑𝑣 = 2
𝑑𝑥 ⇒ 𝑣 = ∫ 𝑑𝑥 =
⏞ − √1 − 𝑥 2 .
√1−𝑥 √1−𝑥 2
Then the integral becomes
𝑥 sin−1 𝑥 𝑥
∫ 𝑑𝑥 = ∫ ⏟sin−1 𝑥 ( 𝑑𝑥) = ∫ sin−1 𝑥 𝑑 (−√1 − 𝑥 2 )
√1 − 𝑥 2 𝑢 ⏟√1 − 𝑥 2 ⏟
𝑑𝑣 ∫ 𝑢𝑑𝑣

= −(sin−1 𝑥 )√1 − 𝑥 2 − ∫ −√1 − 𝑥 2 𝑑 (sin−1 𝑥 )

 Page 66 of 125

1
= −(sin−1 𝑥 )√1 − 𝑥 2 + ∫ √1 − 𝑥 2 ( 𝑑𝑥)
√1 − 𝑥2
= − sin−1 𝑥 √1 − 𝑥 2 + ∫ 1𝑑𝑥 = − sin−1 𝑥 √1 − 𝑥 2 + 𝑥 + 𝐶.
Method 2: Using substitution
We let 𝑥 = sin 𝜃 so that 1 − 𝑥 2 = 1 − sin2 𝜃 = cos2 𝜃 (NOT 𝑦 = 1 − 𝑥 2 or 𝑥 =
𝑑𝑥
cos 𝜃!!). Then = cos 𝜃 and the integral becomes
𝑑𝜃

𝑥 sin−1 𝑥 sin 𝜃 [sin−1 (sin 𝜃)]

∫ 𝑑𝑥 = ∫ (cos 𝜃 𝑑𝜃) = ∫ 𝜃 sin 𝜃 𝑑𝜃
√1 − 𝑥2 √1 − sin2 𝜃

= ∫ 𝜃(sin 𝜃 𝑑𝜃) = ∫ 𝜃𝑑(− cos 𝜃) = −𝜃 cos 𝜃 − ∫(− cos 𝜃)𝑑𝜃

= −𝜃 cos 𝜃 + sin 𝜃 + 𝐶 = −(sin−1 𝑥 ) (√1 − 𝑥 2 ) + ⏟

𝑥 + 𝐶.
sin(sin−1 𝜃 )
 Page 67 of 125

Reduction Formula
How to integrate the following integrals

∫ 𝑥 10 𝑒 𝑥 𝑑𝑥 , ∫ cos8 𝑥 𝑑𝑥 , ∫(ln 𝑥)20 𝑑𝑥 ?

For example, we wish to compute the first integral,

∫ 𝑥 10 𝑒 𝑥 𝑑𝑥,

 One can use integration by parts to transform the integral into simpler one
(Eliminate 𝑥 10 in the integrand through differentiation).

 It cannot be done in one shot (See Example 17)!

 Page 68 of 125

One can use integration by parts 10 times to find out the answer, i.e.,

∫ 𝑥 10 𝑒 𝑥 𝑑𝑥 = ∫ 𝑥 10 𝑑𝑒 𝑥 = 𝑥 10 𝑒 𝑥 − ∫ 𝑒 𝑥 𝑑𝑥 10 = 𝑥 10 𝑒 𝑥 − 10 ∫ 𝑥 9 𝑒 𝑥 𝑑𝑥

= 𝑥 10 𝑒 𝑥 − 10 ∫ 𝑥 9 𝑑𝑒 𝑥 = 𝑥 10 𝑒 𝑥 − 10 (𝑥 9 𝑒 𝑥 − ∫ 𝑒 𝑥 𝑑𝑥 9 )

= 𝑥 10 𝑒 𝑥 − 10𝑥 9 𝑒 𝑥 + 90 ∫ 𝑥 8 𝑒 𝑥 𝑑𝑥

= 𝑥 10 𝑒 𝑥 − 10𝑥 9 𝑒 𝑥 + 90 ∫ 𝑥 8 𝑑𝑒 𝑥 = 𝑥 10 𝑒 𝑥 − 10𝑥 9 𝑒 𝑥 + 90 (𝑥 8 𝑒 𝑥 − ∫ 𝑒 𝑥 𝑑𝑥 8 )

= 𝑥 10 𝑒 𝑥 − 10𝑥 9 𝑒 𝑥 + 90𝑥 8 𝑒 𝑥 − 720 ∫ 𝑥 7 𝑒 𝑥 𝑑𝑥

= ⋯ (OK, I give up)

 Page 69 of 125

 For these “massive” integrals, it requires repeated use of integration by parts.

This causes a huge computational cost.

 Question: Is there any shortcut?

 Answer: Reduction Formula

Reduction Formula is an integration technique which tries to compute the integrals
which the expressions involving powers of elementary functions (e.g. cosn 𝑥,
(𝑎𝑥 2 + 𝑏𝑥 + 𝑐 )𝑛 , (ln 𝑥 )𝑛 etc.).
 Let 𝐼𝑛 = ∫ cos𝑛 𝑥 𝑑𝑥 , ∫(𝑎𝑥 2 + 𝑏𝑥 + 𝑐 )𝑛 𝑑𝑥 , ∫(ln 𝑥 )𝑛 𝑑𝑥, we first use some
technique (usually integration by parts) to derive the equation (called
recurrence relation or reduction formula) governing 𝐼𝑛 .
 Then we compute our target integral using that recurrence relation derived in
the first step.
 Page 70 of 125

Consider the integral ∫ 𝑥 4 𝑒 𝑥 𝑑𝑥 as an example, we compute this integral as follows:

Step 1: Derive the reduction formula
We let 𝐼𝑛 = ∫ 𝑥 𝑛 𝑒 𝑥 𝑑𝑥, where 𝑛 is any non-negative integer. We use the
integration by parts (with 𝑢(𝑥 ) = 𝑥 𝑛 and 𝑑𝑣 = 𝑒 𝑥 𝑑𝑥) and obtain
𝑑𝑣=𝑒 𝑥 𝑑𝑥
⇒ 𝑣=∫ 𝑒 𝑥 𝑑𝑥=𝑒 𝑥
𝐼𝑛 = ∫ 𝑥⏟𝑛 ⏟
𝑒 𝑥 𝑑𝑥 =
⏞ ∫ 𝑥⏟𝑛 𝑑(𝑒
⏟ 𝑥 ) = 𝑥 𝑛 𝑒 𝑥 − ∫ 𝑒 𝑥 𝑑 (𝑥 𝑛 )
𝑢 𝑑𝑣 𝑢 𝑑𝑣

= 𝑥 𝑛 𝑒 𝑥 − ∫ 𝑒 𝑥 (𝑛𝑥 𝑛−1 )𝑑𝑥 = 𝑥 𝑛 𝑒 𝑥 − 𝑛 ∫ 𝑥 𝑛−1 𝑒 𝑥 𝑑𝑥 = 𝑥 𝑛 𝑒 𝑥 − 𝑛𝐼𝑛−1 .

Hence, we obtain a reduction formula for 𝐼𝑛 :

𝐼𝑛 = 𝑥 𝑛 𝑒 𝑥 − 𝑛𝐼𝑛−1
 Page 71 of 125

Step 2: Use reduction formula to compute the target integral

Note that the target integral is just 𝐼4 = ∫ 𝑥 4 𝑒 𝑥 𝑑𝑥, using the formula with 𝑛 = 4
obtained in Step 1, we then have
𝐼4 = 𝑥 4 𝑒 𝑥 − 4𝐼3 Substitute 𝑛 = 3:
𝐼3 = 𝑥 3 𝑒 𝑥 − 3𝐼2
= 𝑥 4 𝑒 𝑥 − 4(𝑥 3 𝑒 𝑥 − 3𝐼2 )
= 𝑥 4 𝑒 𝑥 − 4𝑥 3 𝑒 𝑥 + 12𝐼2 Substitute 𝑛 = 2:
= 𝑥 4 𝑒 𝑥 − 4𝑥 3 𝑒 𝑥 + 12(𝑥2 𝑒𝑥 − 2𝐼1 ) 𝐼2 = 𝑥 2 𝑒 𝑥 − 2𝐼1

= 𝑥 4 𝑒 𝑥 − 4𝑥 3 𝑒 𝑥 + 12𝑥2 𝑒𝑥 − 24𝐼1 Substitute 𝑛 = 1:

𝐼1 = 𝑥𝑒 𝑥 − 𝐼0
= 𝑥 4 𝑒 𝑥 − 4𝑥 3 𝑒 𝑥 + 12𝑥2 𝑒𝑥 − 24(𝑥𝑒𝑥 − 𝐼0 )

= 𝑥 4 𝑒 𝑥 − 4𝑥 3 𝑒 𝑥 + 12𝑥2 𝑒𝑥 − 24𝑥𝑒 𝑥 + 24 ∫ 𝑥 0 𝑒 𝑥 𝑑𝑥
⏟
∫ 𝑒 𝑥 𝑑𝑥

= 𝑥 4 𝑒 𝑥 − 4𝑥 3 𝑒 𝑥 + 12𝑥2 𝑒𝑥 − 24𝑥𝑒 𝑥 + 24𝑒 𝑥 + 𝐶.

 Page 72 of 125

Example 25
𝑒 𝑒
Derive the reduction formula of 𝐼𝑛 = ∫1 (ln 𝑥 )𝑛 𝑑𝑥 and find ∫1 (ln 𝑥 )3 𝑑𝑥.

Solution:
Step 1: Derive the reduction formula

Using Integration by parts with 𝑢 = (ln 𝑥 )𝑛 and 𝑑𝑣 = 𝑑𝑥 ⇒ 𝑣 = ∫ 𝑑𝑥 = 𝑥, we

have 𝑑 𝑑 ( ln 𝑥 )𝑛 (
𝑑 ln 𝑥 )
𝑛(ln 𝑥 ) =
𝑒 𝑒 𝑑𝑥 𝑑(ln 𝑥) 𝑑𝑥
𝐼𝑛 = ∫ ⏟ ⏟ = [𝑥 (ln 𝑥 )𝑛 ]1𝑒 − ∫ 𝑥𝑑(ln 𝑥 )𝑛
(ln 𝑥 )𝑛 𝑑𝑥 1
1 𝑑𝑣 1 = 𝑛(ln 𝑥 )𝑛−1 ( )
𝑢 𝑥
𝑒
1
= 𝑒 − ∫ 𝑥 [𝑛(ln 𝑥 )𝑛−1 ( )] 𝑑𝑥
1 𝑥
𝑒
= 𝑒 − 𝑛 ∫ (ln 𝑥 )𝑛−1 𝑑𝑥 = 𝑒 − 𝑛𝐼𝑛−1 .
1
 Page 73 of 125

Therefore, we obtain the following reduction formula for 𝐼𝑛

𝐼𝑛 = 𝑒 − 𝑛𝐼𝑛−1

Step 2: Use reduction formula to compute the target integral

𝑒
Note that 𝐼3 = ∫1 (ln 𝑥 )3 𝑑𝑥, using the reduction formula with 𝑛 = 3, we then have
𝑒 𝑛=3
∫ (ln 𝑥 )3 𝑑𝑥 = 𝐼3 =
⏞ 𝑒 − 3𝐼2
1
𝑛=2
⏞ 𝑒 − 3(𝑒 − 2𝐼1 ) = −2𝑒 + 6𝐼1
=
𝑛=1 𝑒
⏞ − 2𝑒 + 6(𝑒 − 1𝐼0 ) = 4𝑒 − 6𝐼0 = 4𝑒 − 6 ∫ (ln 𝑥 )0 𝑑𝑥
=
1
𝑒
= 4𝑒 − 6 ∫ 1𝑑𝑥 = 4𝑒 − 6[𝑥 ]1𝑒 = 4𝑒 − 6𝑒 + 6 = 6 − 2𝑒.
1
 Page 74 of 125

Remark of reduction formula techniques

 As seen in the above 2 examples, the reduction formula can greatly reduce the
computation cost (at least no need to write so many steps). For definite integrals,
the reduction formula provides a nice computation formula which is “computer-
friendly”. Take Example 25 as an example, one can use the recurrent formula
obtained in Step 1 and computer software (say excel, Matlab etc.,) to compute
the integral likes
𝑒
∫ (ln 𝑥 )50 𝑑𝑥 .
1
 Since the reduction formula is derived by using integration by parts formula, one
would like to ask what is the suitable choice of 𝑢(𝑥). Recall that the main
purpose of reduction is to “reduce” the power 𝑛 in the integral. Therefore in
most of the cases, your 𝑢(𝑥) should be chosen to be (? ? )𝑛 (for example, 𝑥 𝑛
for ∫ 𝑥 𝑛 𝑒 𝑥 𝑑𝑥 or (ln 𝑥 )𝑛 in Example 25.
 Page 75 of 125

Example 26
(a) Derive the reduction formula for the integral
1
1
𝐼𝑛 = ∫ 𝑑𝑥 , 𝑛 ≥ 1
0 (1 + 𝑥 2 )𝑛

(b) Hence, compute the integral

1
1
∫ 𝑑𝑥
0 (1 + 𝑥 2 )3

Solution of (a)
1
Using the integration by parts with 𝑢(𝑥 ) = ( and 𝑑𝑣 = 𝑑𝑥 ⇒ 𝑣 = ∫ 𝑑𝑥 = 𝑥,
1+𝑥 2 )𝑛
we have
1 1 1
1 𝑥 1
𝐼𝑛 = ∫ 𝑑𝑥
⏟=[ ] −∫ 𝑥 𝑑( )
(
0 ⏟1 + 𝑥 2 )𝑛
𝑑𝑣
(1 + 𝑥 2 )𝑛
0 0 ⏟ (1 + 𝑥 2 )𝑛

𝑢 =(1+𝑥 2 )−𝑛
 Page 76 of 125

1 1
1 2 −(𝑛+1) 1 𝑥2
= 𝑛 − ∫ 𝑥(−𝑛(1 + 𝑥 ) (2𝑥 )𝑑𝑥) = 𝑛 + 2𝑛 ∫ 𝑑𝑥
2 0 2 (
⏟0 1 + 𝑥 2 )𝑛+1

≠𝐼𝑛+1 or 𝐼𝑛 ‼‼
1 1
1 𝑥2 + 1 1
= 𝑛 + 2𝑛 [∫ 𝑑𝑥 − ∫ 𝑑𝑥 ]
2 0 (1 + 𝑥 2 )𝑛+1
0 (1 + 𝑥 2 )𝑛+1

1 1
1 1 1
= 𝑛 + 2𝑛 [∫ 𝑑𝑥 − ∫ 𝑑𝑥 ]
2 0 (1 + 𝑥 2 )𝑛
0 (1 + 𝑥 2 )𝑛+1

1
= 𝑛 + 2𝑛𝐼𝑛 − 2𝑛𝐼𝑛+1
2
After some rearrangement, we finally obtain
1 1
𝐼𝑛 = 𝑛 + 2𝑛𝐼𝑛 − 2𝑛𝐼𝑛+1 ⇒ 2𝑛𝐼𝑛+1 = 𝑛 + (2𝑛 − 1)𝐼𝑛
2 2
𝑛≠ 0 1 2𝑛 − 1
⇒ 𝐼𝑛+1 =
⏞ 𝑛+1 + 𝐼𝑛 .
2 𝑛 2𝑛
 Page 77 of 125

Solution of (b)
Using the reduction formula obtained in (a), the integral can be computed as
1
1
∫ 𝑑𝑥 = 𝐼3
(
0 1+𝑥
2 )3

𝑛=2 1 2(2) − 1 1 3
=
⏞ 2+1 + 𝐼 = + 𝐼
2 (2) 2(2) 2 16 4 2
𝑛=11 3 1 2(1) − 1 1 3
=
⏞ + ( 1+1 + 𝐼1 ) = + 𝐼1
16 4 2 (1) 2(1) 4 8
1 3 1 1 1 3 −1 1
1 3𝜋
= + ∫ 2
𝑑𝑥 = + [⏟tan 𝑥 ]0 = + .
4 8 0 1+𝑥 4 8 −1 −1
4 32
tan 1−tan 0

Reminder:
When computing the numerical value of the inverse trigonometric function, take the
value within the principal range.
 Page 78 of 125

Some advanced examples about the derivation of reduction formula

Example 27
(a) Derive the reduction formula for the integral
2 +1
𝐼𝑛 = ∫ 𝑥 𝑛 𝑒 𝑥 𝑑𝑥.
(b) Hence, compute the integral
2 +1
𝐼5 = ∫ 𝑥 5 𝑒 𝑥 𝑑𝑥.

IDEA:
We will encounter difficulty in deriving (a) if we use integration by parts with 𝑢(𝑥 ) =
2 2
𝑥 𝑛 and 𝑑𝑣 = 𝑒 𝑥 +1 𝑑𝑥. It is because the function 𝑣 = ∫ 𝑒 𝑥 +1 𝑑𝑥 cannot be
obtained. Here, we need to make some modification so that the integration by parts
can be applied. Note that
𝑦=𝑥 2 +1
𝑥 2 +1
1 1 1 2
∫ 𝑥𝑒 𝑑𝑥 =
⏞ ∫ 𝑒 𝑦 𝑑𝑦 = 𝑒 𝑦 + 𝐶 = 𝑒 𝑥 +1 + 𝐶.
2 2 2
 Page 79 of 125

Solution of (a)
2 +1 2 +1 1 2 +1
Here, we take 𝑢(𝑥 ) = 𝑥 𝑛−1 and 𝑑𝑣 = 𝑥𝑒 𝑥 𝑑𝑥 ⇒ 𝑣 = ∫ 𝑥𝑒 𝑥 𝑑𝑥 = 𝑒 𝑥 .
2
Then we have

𝑛 𝑥 2 +1 𝑛−1 𝑥 2 +1
1 𝑥2 +1
𝑛−1
𝐼𝑛 = ∫ 𝑥 𝑒 𝑑𝑥 = ∫ 𝑥⏟ (𝑥𝑒
⏟ 𝑑𝑥) = ∫ 𝑥⏟ 𝑑 ( 𝑒 )
⏟ 2
𝑢 𝑑𝑣 𝑢
𝑑𝑣

1 𝑛−1 𝑥2 +1 1 𝑥2+1
= 𝑥 𝑒 −∫ 𝑒 𝑑 (𝑥 𝑛−1 )
2 2
1 𝑛−1 𝑥2 +1 𝑛 − 1 2
= 𝑥 𝑒 − ∫ 𝑥 𝑛−2 𝑒 𝑥 +1 𝑑𝑥
2 2
1 𝑛−1 𝑥2 +1 𝑛 − 1
= 𝑥 𝑒 − 𝐼
2 2 𝑛−2
 Page 80 of 125

Solution of (b)
Using the formula with 𝑛 = 5, we have
2 +1 1 2 4
∫ 𝑥 5𝑒 𝑥 𝑑𝑥 = 𝐼5 = 𝑥 4 𝑒 𝑥 +1 − 𝐼3
2 2
𝑛=3 1 1 2 𝑥2+1 2
4 𝑥 2 +1
=
⏞ 𝑥 𝑒 − 2( 𝑥 𝑒 − 𝐼1 )
2 2 2
1 4 𝑥2+1 2 2
= 𝑥 𝑒 − 𝑥 2 𝑒 𝑥 +1 + 2 ∫ 𝑥𝑒 𝑥 +1 𝑑𝑥
2
∫ 𝑥𝑒 𝑥2+1 𝑑𝑥
1 2
= 𝑒 𝑥 +1 +𝐶
2 1 2 +1 2 +1 2 +1
=
⏞ 𝑥 4𝑒 𝑥 − 𝑥 2𝑒 𝑥 + 𝑒𝑥 +𝐶
2
 Page 81 of 125

Example 28
Derive a reduction formula for
𝜋
2
𝑆𝑛 = ∫ sin𝑛 𝑥 𝑑𝑥 ,
0

where 𝑛 = 2, 3, 4, 5, ….
IDEA:
Here one should not use integration by parts with 𝑢(𝑥 ) = sin𝑛 𝑥 and 𝑑𝑣 =
𝑑𝑥 ⇒ 𝑣 = ∫ 𝑑𝑥 = 𝑥 directly since there will be a “undesired function”
coming out after the transformation, i.e.
𝜋 𝜋
𝜋
2 2
⏟ 𝑛 𝑥 𝑑𝑥
∫ sin ⏟= [𝑥 sin𝑛 𝑥 ]02 − ∫ 𝑥𝑑(sin𝑛 𝑥 )
0 𝑢 𝑑𝑣 0
𝜋
𝜋 2
= −𝑛∫ 𝑥 sin𝑛−1 𝑥 cos 𝑥 𝑑𝑥 .
⏟
2 0 extra
 Page 82 of 125

Solution:
𝜋
Let 𝑆𝑛 = ∫0 sin𝑛 𝑥 𝑑𝑥
2

Similar to Example 20, one should use integration by parts in the following way:
𝜋 𝜋
2 2
𝑆𝑛 = ∫ sin𝑛 𝑥 𝑑𝑥 = ∫ ⏟
sin𝑛−1 𝑥 ⏟
(sin 𝑥 𝑑𝑥 )
0 0 𝑢 𝑑𝑣
𝑑𝑣=sin 𝑥𝑑𝑥
⇒𝑣=∫ sin 𝑥𝑑𝑥=− cos 𝑥 𝜋
2
=
⏞ sin𝑛−1 𝑥 ⏟
∫ ⏟ 𝑑 (− cos 𝑥 )
0 𝑢 𝑑𝑣
𝜋
𝜋
2
= [− sin𝑛−1 𝑥 cos 𝑥 ]02 − ∫ − cos 𝑥 𝑑 (sin𝑛−1 𝑥 )
0
𝜋
2
= 0 + ∫ cos 𝑥 [(𝑛 − 1) sin𝑛−2 𝑥 cos 𝑥 𝑑𝑥 ]
0
 Page 83 of 125

𝜋
2
= (𝑛 − 1) ∫ cos2 𝑥 sin𝑛−2 𝑥 𝑑𝑥
0
𝜋
sin2 𝑥+cos2 𝑥=1 2
=
⏞ (𝑛 − 1) ∫ (1 − sin2 𝑥 ) sin𝑛−2 𝑥 𝑑𝑥
0
𝜋 𝜋
2 2
= (𝑛 − 1) ∫ sin𝑛−2 𝑥 𝑑𝑥 − (𝑛 − 1) ∫ sin𝑛 𝑥 𝑑𝑥
0 0

= (𝑛 − 1)𝑆𝑛−2 − (𝑛 − 1)𝑆𝑛
So we have
𝑆𝑛 = (𝑛 − 1)𝑆𝑛−2 − (𝑛 − 1)𝑆𝑛
𝑛−1
⇒ 𝑆𝑛 = 𝑆𝑛−2 .
𝑛
 Page 84 of 125

To compute the exact value of 𝑆𝑛 , one can use this formula repeatedly, we get
replace 𝑛
by 𝑛−2
𝑛−1 𝑛−1 𝑛−3
𝑆𝑛 = 𝑆𝑛−2 =
⏞ ( 𝑆𝑛−4 )
𝑛 𝑛 𝑛−2
replace 𝑛
by 𝑛−4
𝑛 − 1𝑛 − 3 𝑛 − 5
=
⏞ ( 𝑆 )
𝑛 𝑛 − 2 𝑛 − 4 𝑛−6
(𝑛 − 1) × (𝑛 − 3) × … × 5 × 3 × 1
𝑆0 if 𝑛 is even
𝑛 × (𝑛 − 2) × … × 6 × 4 × 2
=⋯=
(𝑛 − 1) × (𝑛 − 3) × … × 6 × 4 × 2
𝑆1 if 𝑛 is odd
{ 𝑛 × (𝑛 − 2) × … × 7 × 5 × 3
Since
𝜋 𝜋 𝜋
𝜋
2 𝜋2 2
0
𝑆0 = ∫ sin 𝑥 𝑑𝑥 = ∫ 𝑑𝑥 = , 𝑆1 = ∫ sin 𝑥 𝑑𝑥 = [− cos 𝑥 ]02 =1
0 0 2 0
𝜋
Therefore, we obtain the formula for 𝑆𝑛 = ∫0 sin𝑛 𝑥 𝑑𝑥 :
2
 Page 85 of 125

(𝑛 − 1) × (𝑛 − 3) × … × 5 × 3 × 1 𝜋
× if 𝑛 is even
𝑛 × (𝑛 − 2) × … × 6 × 4 × 2 2
𝑆𝑛 = .
(𝑛 − 1) × (𝑛 − 3) × … × 6 × 4 × 2
if 𝑛 is odd
{ 𝑛 × (𝑛 − 2) × … × 7 × 5 × 3
This formula is called Wallis’s formula.
Remark:
𝜋
Using similar method, one can obtain the similar formula for 𝐶𝑛 = ∫0 cos𝑛 𝑥 𝑑𝑥.
2

(Left as exercise)

(𝑛 − 1) × (𝑛 − 3) × … × 5 × 3 × 1 𝜋
𝜋 ( ) if 𝑛 is even
2 𝑛 × (𝑛 − 2) × … × 6 × 4 × 2 2
𝐶𝑛 = ∫ cos𝑛 𝑥 𝑑𝑥 =
0 (𝑛 − 1) × (𝑛 − 3) × … × 6 × 4 × 2
if 𝑛 is odd
{ 𝑛 × (𝑛 − 2) × … × 7 × 5 × 3
 Page 86 of 125

Example 29
(a) Derive the reduction formula for the following integral
𝐼𝑛 = ∫ sec 𝑛 𝑥 𝑑𝑥
(b) Hence, compute the following integral
1
∫ 𝑑𝑥
(1 − 𝑥 2 )3
Solution of (a)
Using similar tactic as in Example 28, we use integration of parts with 𝑢(𝑥 ) =
sec 𝑛−2 𝑥 and 𝑑𝑣 = sec 2 𝑥 𝑑𝑥 ⇒ 𝑣 = ∫ sec 2 𝑥 𝑑𝑥 = tan 𝑥 and obtain

𝐼𝑛 = ∫ sec 𝑛 𝑥 𝑑𝑥 = ∫ ⏟
sec 𝑛−2 𝑥 (⏟sec 2 𝑥 𝑑𝑥 ) = ∫ ⏟
sec 𝑛−2 𝑥 ⏟
𝑑(tan 𝑥 )
𝑢 𝑑𝑣 𝑢 𝑑𝑣

= sec 𝑛−2 𝑥 tan 𝑥 − ∫ tan 𝑥 𝑑(sec 𝑛−2 𝑥 )

 Page 87 of 125

= sec 𝑛−2 𝑥 tan 𝑥 − ∫ tan 𝑥 ⏟

[(𝑛 − 2) sec 𝑛−3 𝑥 (sec 𝑥 tan 𝑥 )𝑑𝑥 ]
𝑑 𝑑 (sec𝑛−2 𝑥)𝑑 (sec 𝑥)
(sec𝑛−2 𝑥)=
𝑑𝑥 𝑑 (sec 𝑥) 𝑑𝑥

= sec 𝑛−2 𝑥 tan 𝑥 − (𝑛 − 2) ∫ sec 𝑛−2 𝑥 tan2 𝑥 𝑑𝑥

1+tan2 𝑥=sec2 𝑥
=
⏞ sec 𝑛−2 𝑥 tan 𝑥 − (𝑛 − 2) ∫ sec 𝑛−2 𝑥 (sec 2 𝑥 − 1)𝑑𝑥

= sec 𝑛−2 𝑥 tan 𝑥 − (𝑛 − 2) ∫ sec 𝑛 𝑥 𝑑𝑥 + (𝑛 − 2) ∫ sec 𝑛−2 𝑥 𝑑𝑥

= sec 𝑛−2 𝑥 tan 𝑥 − (𝑛 − 2)𝐼𝑛 + (𝑛 − 2)𝐼𝑛−2 .

Summing up, we obtain
𝐼𝑛 = sec 𝑛−2 𝑥 tan 𝑥 − (𝑛 − 2)𝐼𝑛 + (𝑛 − 2)𝐼𝑛−2
1 𝑛−2
𝑛−2
⇒ 𝐼𝑛 = sec 𝑥 tan 𝑥 + 𝐼 .
𝑛−1 𝑛 − 1 𝑛−2
 Page 88 of 125

Solution of (b)
As a starting point, one may try the simple substitution 𝑦 = 1 − 𝑥 2 as the
first trial. The integral is then becomes
1 1 1
∫ 𝑑𝑥 = − ∫ 𝑑𝑦
(1 − 𝑥 2 )3 2 𝑦 √1 − 𝑦
3

which is complicated to integrate. Hence, we should try trigonometric

substitution as an alternative.
𝑑𝑥
We let 𝑥 = sin 𝜃. Then = cos 𝜃 and the integral becomes
𝑑𝜃

1 1 1 5
∫ 𝑑𝑥 = ∫ 3
(cos 𝜃 𝑑𝜃 ) = ∫ 𝑑𝜃 = ∫ sec 𝜃 𝑑𝜃.
(1 − 𝑥 2 )3 cos5 𝜃
⏟− sin2 𝜃 )
(1
=cos2 𝜃
 Page 89 of 125

Using the reduction formula obtained in (a), the above integral is found to be
1𝑛=5 5−2
5 5−2
∫ sec 𝜃 𝑑𝜃 = 𝐼5 =
⏞ sec 𝜃 tan 𝜃 + 𝐼
5−1 5−1 3
1 3
3
= sec 𝜃 tan 𝜃 + 𝐼3
4 4
𝑛=3 1 3 1 3−2
3 3−2
=
⏞ sec 𝜃 tan 𝜃 + ( sec 𝜃 tan 𝜃 + 𝐼)
4 4 3−1 3−1 1
1 3
3 3
= sec 𝜃 tan 𝜃 + sec 𝜃 tan 𝜃 + ∫ sec 𝜃 𝑑𝜃
4 8 8
1 3
3 3
= sec 𝜃 tan 𝜃 + sec 𝜃 tan 𝜃 + ln|sec 𝜃 + tan 𝜃| + 𝐶
4 8 8
1 𝑥
sec 𝜃= , tan 𝜃=
√1−𝑥 2 √1−𝑥 2 1 𝑥 3 𝑥 3 1+𝑥
=
⏞ + + ln | | + 𝐶.
(
4 1−𝑥 2 )2 81 − 𝑥 2 8 √1 − 𝑥 2
 Page 90 of 125

Integration of rational function: Method of Partial Fractions

In this section, we shall investigate the integral of the rational function 𝑅(𝑥) which is
the quotient of two polynomials:
𝑃(𝑥)
∫ 𝑅(𝑥 )𝑑𝑥 = ∫ 𝑑𝑥
𝑄(𝑥)
where both 𝑃(𝑥) and 𝑄(𝑥) are polynomials.
When we integrate the rational function, it is always convenient for us to first split
(decompose) the function into the sum of simpler (rational) functions. This could be
done by method of partial fractions.
𝑥 3 + 5𝑥 2 + 2𝑥 − 2 1 1 𝑥−2
∫ 𝑑𝑥 = ∫ 𝑑𝑥 + ∫ 𝑑𝑥 − ∫ 2 𝑑𝑥
(𝑥 − 1)(2𝑥 + 1)(𝑥 2 + 1) 𝑥−1 2𝑥 + 1 𝑥 +1
 Page 91 of 125

A quick review on the method of partial fraction

The whole process is divided into three steps can be summarized by the following
flow chart:
Step 1 The function Step 2
is proper Factorize the denominator of the
Determine whether a
given rational function proper rational function and choose
is proper or improper. the decomposition of the function.
(See Appendix A for details)
The function
is improper
Step 3
Extra
Determine the unknown in the
Perform long division to
proposed decomposition (using
rewrite the function into the
sum of a polynomial and a substitution method or
proper rational function comparing coefficients)
 Page 92 of 125

Example 30
Compute the integral
1
∫ 𝑑𝑥
(𝑥 − 3)(2𝑥 + 1)(𝑥 − 1)
Solution:
We first decompose the function using the method partial fraction. Note that the
function is proper (degree of numerator = 0 < degree of denominator = 3) and the
factors in the denominator are linear and distinct, so we propose the following
decomposition:
1 𝐴 𝐵 𝐶
= + + .
(𝑥 − 3)(2𝑥 + 1)(𝑥 − 1) 𝑥 − 3 2𝑥 + 1 𝑥 − 1
By taking the common denominator, we obtain the following equation:
1 = 𝐴(2𝑥 + 1)(𝑥 − 1) + 𝐵(𝑥 − 3)(𝑥 − 1) + 𝐶 (𝑥 − 3)(2𝑥 + 1).
1
Substitute 𝑥 = 1, we obtain 1 = 𝐶 (−2)(3) ⇒ 𝐶 = − .
6
 Page 93 of 125

1 7 3 4
Substitute 𝑥 = − , we obtain 1 = 𝐵 (− ) (− ) ⇒ 𝐵 = .
2 2 2 21
1
Substitute 𝑥 = 3, we obtain 1 = 𝐴(7)(2) ⇒ 𝐴 = .
14

Hence, we have
1 1 4 1
= + −
(𝑥 − 3)(2𝑥 + 1)(𝑥 − 1) 14(𝑥 − 3) 21(2𝑥 + 1) 6(𝑥 − 1)
Then we can compute the integral as
1
∫ 𝑑𝑥
(𝑥 − 3)(2𝑥 + 1)(𝑥 − 1)
1 1 4 1 1 1
= ∫ 𝑑𝑥 + ∫ 𝑑𝑥 − ∫ 𝑑𝑥
14 𝑥 − 3 21 2𝑥 + 1 6 𝑥−1
1 2 1
= ln 𝑥 − 3 + ln 2𝑥 + 1 − ln|𝑥 − 1| + 𝐶.
| | | |
14 21 6
 Page 94 of 125

Example 31
Compute the integrals
1 1
∫ 2
𝑑𝑥 , ∫ 2
𝑑𝑥
𝑥 + 4𝑥 − 12 𝑥 + 4𝑥 + 12
Solution:
For the first integral, one can rewrite the integral as
1 1
∫ 2 𝑑𝑥 = ∫ 𝑑𝑥
𝑥 + 4𝑥 − 12 (𝑥 + 6)(𝑥 − 2)
Then we decompose the function using the method of partial fractions as follows:
1 𝐴 𝐵
= + ⇒ 1 = 𝐴(𝑥 − 2) + 𝐵(𝑥 + 6).
(𝑥 + 6)(𝑥 − 2) 𝑥 + 6 𝑥 − 2
1
Substitute 𝑥 = 2, we get 1 = 8𝐵 ⇒ 𝐵 = .
8
1
Substitute 𝑥 = −6, we get 1 = −8𝐴 ⇒ 𝐴 = − .
8

Thus the integral can be computed as

 Page 95 of 125

1 1 1 1 1
∫ 𝑑𝑥 = − ∫ 𝑑𝑥 + ∫ 𝑑𝑥
(𝑥 + 6)(𝑥 − 2) 8 𝑥+6 8 𝑥−2
1 1
= − ln 𝑥 + 6 + ln|𝑥 − 2| + 𝐶.
| |
8 8
For the second integral, note that the denominator cannot be factorized, we can
compute the integral using the technique in Example 15 of Chapter 1 (pp. 27-28) :
1 1 1 1
∫ 2 𝑑𝑥 = ∫ 𝑑𝑥 = ∫ 𝑑𝑥
𝑥 + 4𝑥 + 12 (𝑥 + 2)2 + 8 8 1 (𝑥 + 2)2 + 1
8

1 1 1 1 −1
1 2
= ∫ 𝑑𝑥 = ( tan ( 𝑥 + ))
8 1 2 2 8 1 √8 √8
( 𝑥+ ) +1
√8 √8 √8
1 −1
1 2
= tan ( 𝑥 + ) + 𝐶.
√8 √8 √8
 Page 96 of 125

Remark of Example 31
 In this example, we see that there are two totally different ways to encounter
𝐴
the integral of the form ∫ 2 𝑑𝑥: Method of partial fractions and
𝑎𝑥 +𝑏𝑥+𝑐
Technique in Example 15 of Chapter 1.

 This depends on whether the denominator can be further factorized into two
linear functions.

 To check whether a given quadratic function 𝑎𝑥 2 + 𝑏𝑥 + 𝑐 can be factorized or

not, one can use the discriminant Δ = 𝑏 2 − 4𝑎𝑐:
 If Δ ≥ 0, the equation 𝑎𝑥 2 + 𝑏𝑥 + 𝑐 = 0 has at least one real root 𝛼, 𝛽.
Then the function can be factorized into 𝑎(𝑥 − 𝛼)(𝑥 − 𝛽).
 If Δ < 0, the equation 𝑎𝑥 2 + 𝑏𝑥 + 𝑐 = 0 has no real roots and the given
function cannot be factorized.
 Page 97 of 125

Example 32
Computed the integral
5𝑥 3 − 12𝑥 2 + 5𝑥 − 4
∫ 𝑑𝑥.
(2𝑥 + 1)(𝑥 − 1)3
Solution
We first decompose the rational function using the method of partial fraction. Since
the factorization of denominator involve repeated factor, we should propose the
following decomposition:
5𝑥 3 − 12𝑥 2 + 5𝑥 − 4 𝐴 𝐵 𝐶 𝐷
= + + + .
(2𝑥 + 1)(𝑥 − 1) 3 2𝑥 + 1 𝑥 − 1 (𝑥 − 1) 2 (𝑥 − 1)3

Multiplying the common denominator on both sides, we then obtain

𝐴(𝑥 − 1)3 + 𝐵(𝑥 − 1)2 (2𝑥 + 1) + 𝐶 (𝑥 − 1)(2𝑥 + 1) + 𝐷(2𝑥 + 1)
= 5𝑥 3 − 12𝑥 2 + 5𝑥 − 4.
1 3 3 1 1 5
Substitute 𝑥 = − , we get 𝐴 (− ) = 5 (− ) − 12 ( ) − − 4 ⇒ 𝐴 = 3.
2 2 8 4 2
 Page 98 of 125

Substitute 𝑥 = 1, we get 3𝐷 = 5 − 12 + 5 − 4 ⇒ 𝐷 = −2.

Compare the coefficient of 𝑥 3 : 𝐴 + 2𝐵 = 5 ⇒ 𝐵 = 1.
Compare the constant term: −𝐴 + 𝐵 − 𝐶 + 𝐷 = −4 ⇒ 𝐶 = 0.
Then we have
5𝑥 3 − 12𝑥 2 + 5𝑥 − 4 3 1 0 2
= + + −
(2𝑥 + 1)(𝑥 − 1)3 2𝑥 + 1 𝑥 − 1 (𝑥 − 1)2 (𝑥 − 1)3
The integral can be computed as
5𝑥 3 − 12𝑥 2 + 5𝑥 − 4 1 1 1
∫ 𝑑𝑥 = 3 ∫ 𝑑𝑥 + ∫ 𝑑𝑥 − 2 ∫ 𝑑𝑥
(2𝑥 + 1)(𝑥 − 1)3 2𝑥 + 1 𝑥−1 ⏟ (𝑥 − 1)3

=∫(𝑥−1)−3 𝑑𝑥
3
= ln|2𝑥 + 1| + ln|𝑥 − 1| + (𝑥 − 1)−2 + 𝐶.
2
 Page 99 of 125

Example 33
Compute the integral
6𝑥 2 − 3𝑥 + 7
∫ 2
𝑑𝑥.
(4𝑥 + 1)(𝑥 + 4)
Solution:
We use the method of partial fractions to decompose the rational function. Since
the denominator consists of a quadratic factor 𝑥 2 + 4, so we shall try the following
decomposition:
6𝑥 2 − 3𝑥 + 7 𝐴 𝐵𝑥 + 𝐶
= +
(4𝑥 + 1)(𝑥 2 + 4) 4𝑥 + 1 𝑥 2 + 4
⇒ 𝐴(𝑥 2 + 4) + (4𝑥 + 1)(𝐵𝑥 + 𝐶 ) = 6𝑥 2 − 3𝑥 + 7.
1 65 1 2 1
Substitute 𝑥 = − , we get 𝐴 = 6 (− ) − 3 (− ) + 7 ⇒ 𝐴 = 2.
4 16 4 4

Compare the coefficient of 𝑥 2 : 𝐴 + 4𝐵 = 6 ⇒ 𝐵 = 1.

Compare the constant term: 4𝐴 + 𝐶 = 7 ⇒ 𝐶 = −1.
 Page 100 of 125

Thus we obtain
6𝑥 2 − 3𝑥 + 7 2 𝑥−1
= + .
(4𝑥 + 1)(𝑥 2 + 4) 4𝑥 + 1 𝑥 2 + 4
The integral can be computed as
6𝑥 2 − 3𝑥 + 7 1 𝑥−1
∫ 𝑑𝑥 = 2 ∫ 𝑑𝑥 + ∫ 2 𝑑𝑥 … … (∗)
(4𝑥 + 1)(𝑥 2 + 4) 4𝑥 + 1 𝑥 +4
The first term can be computed as
1 1
∫𝑎𝑥+𝑏𝑑𝑥=𝑎 ln|𝑎𝑥+𝑏|+𝐶
1 1
∫ 𝑑𝑥 =
⏞ ln|4𝑥 + 1| + 𝐶.
4𝑥 + 1 4

The second term can be computed as

𝑥−1 𝑥 1
∫ 2 𝑑𝑥 = ∫ 2 𝑑𝑥 − ∫ 2 𝑑𝑥
𝑥 +4 𝑥 +4 𝑥 +4
 Page 101 of 125

𝑦=𝑥 2 +4
𝑑𝑦
⇒ =2𝑥
𝑑𝑥 1
1 1 1
=
⏞ ∫ 𝑑𝑦 − ∫ 2 𝑑𝑥
2 𝑦 4 𝑥
+1
4
1 1 1
= ln|𝑦| − ∫ 𝑑𝑥
2 4 𝑥 2
( ) +1
2
1 2
1 −1
𝑥
= ln|𝑥 + 4| − tan + 𝐶.
2 2 2
Combing the result, we have
6𝑥 2 − 3𝑥 + 7 1 1 2
1 −1
𝑥
∫ 𝑑𝑥 = ln|4𝑥 + 1| + ln|𝑥 + 4| − tan + 𝐶.
(4𝑥 + 1)(𝑥 2 + 4) 2 2 2 2
 Page 102 of 125

Example 34
Compute the integral
1
6𝑥 2 + 9𝑥 + 19
∫ 𝑑𝑥
( )( 2
0 𝑥 + 1 𝑥 + 6𝑥 + 13 )
Solution:
We use the method of partial fractions to decompose the rational function. Since
the factorization of the denominator consists of a quadratic factor (which cannot be
factorized further), we shall propose the following decomposition:
6𝑥 2 + 9𝑥 + 19 𝐴 𝐵𝑥 + 𝐶
= +
(𝑥 + 1)(𝑥 2 + 6𝑥 + 13) 𝑥 + 1 𝑥 2 + 6𝑥 + 13
⇒ 𝐴(𝑥 2 + 6𝑥 + 13) + (𝐵𝑥 + 𝐶 )(𝑥 + 1) = 6𝑥 2 + 9𝑥 + 19.
Substitute 𝑥 = −1, we have 8𝐴 = 16 ⇒ 𝐴 = 2.
Compare the coefficient of 𝑥 2 : 𝐴 + 𝐵 = 6 ⇒ 𝐵 = 4.
Compare the constant term: 13𝐴 + 𝐶 = 19 ⇒ 𝐶 = −7.
 Page 103 of 125

Hence, we can compute the integral as

1 1 1
6𝑥 2 + 9𝑥 + 19 1 4𝑥 − 7
∫ 𝑑𝑥 = 2 ∫ 𝑑𝑥 + ∫ 2 𝑑𝑥 … … (∗)
0 (𝑥 + 1)( 𝑥 2 + 6𝑥 + 13)
0 𝑥 + 1 0 𝑥 + 6𝑥 + 13
The first term can be computed as
1
1
∫ 𝑑𝑥 = [ln|𝑥 + 1|]10 = ln 2.
0 𝑥+1

The second term can be computed as

1 1 1
4𝑥 − 7 4𝑥 + 12 19
∫ 2 𝑑𝑥 = ∫ 2 𝑑𝑥 − ∫ 2 𝑑𝑥
0 𝑥 + 6𝑥 + 13 0 𝑥 + 6𝑥 + 13 0 𝑥 + 6𝑥 + 13
𝑦=𝑥 2 +6𝑥+13
𝑑𝑦
⇒ =2𝑥+6 20 1
𝑑𝑥 2 1
=
⏞ ∫ 𝑑𝑦 − 19 ∫ 𝑑𝑥
13 𝑦 0 (𝑥 + 3) 2+4

20
19 1 1 19 1 1
= 2[ln 𝑦]13 − ∫ 𝑑𝑥 = 2[ ln 20 − ln 13] − ∫ 𝑑𝑥
4 0 (𝑥 + 3)2 4 0 𝑥+3 2
+1 ( 2 ) +1
4
 Page 104 of 125

1
19 𝑥 + 3
= 2[ln 20 − ln 13] − [ tan−1 ]
2 2 0
19 −1
19 −1
3
= 2[ln 20 − ln 13] − tan 2 + tan .
2 2 2
Combining the result, we finally get
1
6𝑥 2 + 9𝑥 + 19
∫ 𝑑𝑥
0 (𝑥 + 1)( 𝑥 2 + 6𝑥 + 13)

19 −1
19 −1
3
= 2 ln 2 + 2[ln 20 − ln 13] − tan 2 + tan
2 2 2
1600 19 −1
19 −1
3
= ln ( ) − tan 2 + tan .
169 2 2 2
 Page 105 of 125

Example 35 (Integral of improper rational function)

Compute the integral
3𝑥 4 + 3𝑥 3 − 5𝑥 2 + 𝑥 − 1
∫ 2
𝑑𝑥
𝑥 +𝑥−2
Solution:
Note that the integrand is an improper rational function, one has to
decompose the function into a sum of polynomial and a proper rational
function before using the method of partial fractions.
We first perform the long division (dividing the numerator by the denominator) and
rewrite the integrand as
3𝑥 4 + 3𝑥 3 − 5𝑥 2 + 𝑥 − 1 (𝑥 2 + 𝑥 − 2)(3𝑥 2 + 1) + 1
∫ 𝑑𝑥 = ∫ 𝑑𝑥
𝑥2 + 𝑥 − 2 𝑥2 + 𝑥 − 2
1
= ∫(3𝑥 2 + 1)𝑑𝑥 + ∫ 2 𝑑𝑥 … … (∗)
⏟ ⏟𝑥 + 𝑥 − 2
polynomial proper rational function
 Page 106 of 125

The first integral can be computed as

3
𝑥
∫(3𝑥 2 + 1)𝑑𝑥 = 3 ( ) + 𝑥 + 𝐶 = 𝑥 3 + 𝑥 + 𝐶.
3
We proceed to compute the second integral using the method of partial fraction.
Note that
1 1 1 1 1 1
∫ 2
𝑑𝑥 = ∫ 𝑑𝑥 = ⋯ = ∫ 𝑑𝑥 − ∫ 𝑑𝑥
𝑥 +𝑥−2 (𝑥 + 2)(𝑥 − 1) 3 𝑥−1 3 𝑥+2
1 1
= ln 𝑥 − 1 − ln|𝑥 + 2| + 𝐶.
| |
3 3
Hence,
3𝑥 4 + 3𝑥 3 − 5𝑥 2 + 𝑥 − 1 3
1 1
∫ 2
𝑑𝑥 = 𝑥 + 𝑥 + ln 𝑥 − 1 − ln|𝑥 + 2| + 𝐶.
| |
𝑥 +𝑥−2 3 3
 Page 107 of 125

Example 36
Compute the integral
9𝑥 3 − 46𝑥 2 + 102𝑥 − 9
∫ 𝑑𝑥.
(3𝑥 + 1)2 (𝑥 2 − 6𝑥 + 10)
Solution
Note that the denominator involves a repeated factor (3𝑥 + 1) and a quadratic
factor (𝑥 2 − 6𝑥 + 10), we shall decompose the rational function by using the
following decomposition:
9𝑥 3 − 46𝑥 2 + 102𝑥 − 9 𝐴 𝐵 𝐶𝑥 + 𝐷
= + + .
(3𝑥 + 1)2 (𝑥 2 − 6𝑥 + 10) 3𝑥 + 1 (3𝑥 + 1)2 (𝑥 2 − 6𝑥 + 10)
⇒ 9𝑥 3 − 46𝑥 2 + 102𝑥 − 9
= 𝐴(3𝑥 + 1)(𝑥 2 − 6𝑥 + 10) + 𝐵(𝑥 2 − 6𝑥 + 10) + (𝐶𝑥 + 𝐷 )(3𝑥 + 1)2 .
1 436 109
Substitute 𝑥 = − , we get − = 𝐵 ⇒ 𝐵 = −4.
3 9 9

So 𝐴(3𝑥 + 1)(𝑥 2 − 6𝑥 + 10) + (𝐶𝑥 + 𝐷 )(3𝑥 + 1)2

 Page 108 of 125

= 9𝑥 3 − 46𝑥 2 + 102𝑥 − 9 + 4(𝑥 2 − 6𝑥 + 10) = 9𝑥 3 − 42𝑥 2 + 78𝑥 + 31

1
= (9𝑥 2 − 45𝑥 + 93) (𝑥 + ) = (3𝑥 2 − 15𝑥 + 31)(3𝑥 + 1) by long division
3

Thus 𝐴(𝑥 2 − 6𝑥 + 10) + (𝐶𝑥 + 𝐷 )(3𝑥 + 1) = 3𝑥 2 − 15𝑥 + 31.

1 109 109
Substitute 𝑥 = − , we get 𝐴= ⇒ 𝐴 = 3;
3 9 3

Compare the coefficient of 𝑥 2 : 𝐴 + 3𝐶 = 3 ⇒ 𝐶 = 0;

Compare the constant term: 10𝐴 + 𝐷 = 31 ⇒ 𝐷 = 1.
9𝑥 3 − 46𝑥 2 + 102𝑥 − 9
∫ 𝑑𝑥
(3𝑥 + 1)2 (𝑥 2 − 6𝑥 + 10)
3 4 1
=∫ 𝑑𝑥 − ∫ 𝑑𝑥 + ∫ 2 𝑑𝑥
3𝑥 + 1 (3𝑥 + 1)2 𝑥 − 6𝑥 + 10
4 1
= ln|3𝑥 + 1| + +∫ 𝑑𝑥
3(3𝑥 + 1) (𝑥 − 3)2 + 1
4
= ln|3𝑥 + 1| + + tan−1 (𝑥 − 3) + 𝐶.
3(3𝑥 + 1)
 Page 109 of 125

Appendix A – Various types of decomposition pattern using the method of partial

fraction.
The following table summarizes the proposed decomposition under different type of
factorization of the denominator:

Scenario Proposed decomposition

If 𝑄(𝑥) can be factorized into a 𝐴1 𝐴2 𝐴𝑛
𝑅(𝑥 ) = + +⋯ .
product of distinct linear factor: 𝑎1 𝑥 + 𝑏1 𝑎2 𝑥 + 𝑏2 𝑎𝑛 𝑥 + 𝑏𝑛
𝑃 (𝑥 )
𝑅 (𝑥 ) =
(𝑎1 𝑥 + 𝑏1 )(𝑎2 𝑥 + 𝑏2 ) … (𝑎𝑛 𝑥 + 𝑏𝑛 )
If there is some repeated linear factors 𝐴1 𝐴2 𝐴𝑛
𝑅 (𝑥 ) = + + ⋯ +
in the factorization of 𝑄(𝑥): 𝑎𝑥 + 𝑏 (𝑎𝑥 + 𝑏)2
⏟ (𝑎𝑥 + 𝑏)𝑛
for (𝑎𝑥+𝑏)𝑛
𝑃(𝑥) 𝐵
𝑅 (𝑥 ) = +
(𝑎𝑥 + 𝑏)𝑛 (𝑐𝑥 + 𝑑) ⏟+𝑑
𝑐𝑥
for (𝑐𝑥+𝑑)
 Page 110 of 125

If there is quadratic factor (cannot be degree 1

factorized, e.g. 𝑥 2 + 4) in the ⏞

𝐴𝑥 + 𝐵 𝐶
𝑅(𝑥 ) = 2 +
factorization of 𝑄(𝑥): 𝑎𝑥 + 𝑏𝑥 + 𝑐 𝑑𝑥 + 𝑒
⏟
𝑃(𝑥 ) degree 2
𝑅(𝑥 ) =
(𝑎𝑥 2 + 𝑏𝑥 + 𝑐)(𝑑𝑥 + 𝑒)
If there is some repeated quadratic 𝐴1 𝑥 + 𝐵1 𝐴2 𝑥 + 𝐵2
+ +⋯
factor which cannot be further 𝑎𝑥 2 + 𝑏𝑥 + 𝑐 (𝑎𝑥 2 + 𝑏𝑥 + 𝑐)2
𝑅 (𝑥 ) =
𝐴𝑛 𝑥 + 𝐵𝑛
factorized of 𝑔(𝑥): +
(𝑎𝑥 2 + 𝑏𝑥 + 𝑐)𝑛
⏟
𝑃(𝑥) for (𝑎𝑥 2 +𝑏𝑥+𝑐)𝑛
𝑅(𝑥 ) =
(𝑎𝑥2 + 𝑏𝑥 + 𝑐)𝑛 (𝑑𝑥 + 𝑒) 𝐶
+
𝑑𝑥 + 𝑒
 Page 111 of 125

Appendix B – Brief Table for Integration

𝑥 𝑎+1 2. ∫ 𝑒 𝑥 𝑑𝑥 = 𝑒 𝑥 + 𝐶,
1. ∫ 𝑥 𝑎 𝑑𝑥 = { 𝑎 + 1 if 𝑎 ≠ −1
𝑥
𝑎
∫ 𝑎 𝑥 𝑑𝑥 = + 𝐶, 𝑎>0
ln |𝑥| if 𝑎 = −1 ln 𝑎
3. ∫ sin 𝑥 𝑑𝑥 = − cos 𝑥 + 𝐶 4. ∫ cos 𝑥 𝑑𝑥 = sin 𝑥 + 𝐶

5. ∫ tan 𝑥 𝑑𝑥 = ln|sec 𝑥 | + 𝐶 6. ∫ cot 𝑥 𝑑𝑥 = − ln|csc 𝑥 | + 𝐶

7. ∫ sec 2 𝑥 𝑑𝑥 = tan 𝑥 + 𝐶 8. ∫ csc 2 𝑥 𝑑𝑥 = − cot 𝑥 + 𝐶

9. ∫ sec 𝑥 𝑑𝑥 = ln|sec 𝑥 + tan 𝑥 | + 𝐶 10. ∫ csc 𝑥 𝑑𝑥

= − ln|csc 𝑥 + cot 𝑥 | + 𝐶
11. ∫ sec 𝑥 tan 𝑥 𝑑𝑥 = sec 𝑥 + 𝐶 12. ∫ cot 𝑥 csc 𝑥 𝑑𝑥 = − csc 𝑥 + 𝐶
1 −1
1
13. ∫ 2
𝑑𝑥 = tan 𝑥+𝐶 14. ∫ 𝑑𝑥 = sin−1 𝑥 + 𝐶.
1+𝑥 √1 − 𝑥 2
 Page 112 of 125

Review Examples on integration by parts

Review Example 1
Compute the integral

∫(2𝑥 + 3)𝑒 4𝑥 𝑑𝑥

Solution:
Using integration by parts formula with 𝑢(𝑥 ) = 2𝑥 + 3 and 𝑑𝑣 = 𝑒 4𝑥 𝑑𝑥
1
⇒ 𝑣 = ∫ 𝑒 4𝑥 𝑑𝑥 = 𝑒 4𝑥 , we obtain
4
4𝑥
1 4𝑥 1 4𝑥
(2𝑥 + 3) 𝑒⏟ 𝑑𝑥 = ⏟
∫⏟ (2𝑥 + 3) ( 𝑒 ) − ∫ 𝑒 ⏟ 𝑑 (2𝑥 + 3)
⏟4 ⏟
4
𝑢 𝑑𝑣 𝑢 𝑑𝑢
𝑣 𝑣
𝑑
(2𝑥+3)=2
𝑑𝑥
⇒𝑑 (2𝑥+3)=2𝑑𝑥
1 4𝑥
1 4𝑥
1 4𝑥
1 4𝑥
=
⏞ (2𝑥 + 3)𝑒 − ∫ 𝑒 𝑑𝑥 = (2𝑥 + 3)𝑒 − 𝑒 + 𝐶.
4 2 4 8
 Page 113 of 125

Review Example 2
Compute the integral

∫ √𝑥 (ln 𝑥 )2 𝑑𝑥

Solution:

We use the integration by parts with 𝑢 = (ln 𝑥 )2 and 𝑑𝑣 = √𝑥𝑑𝑥

1 3
2
⇒ 𝑣 = ∫ √𝑥𝑑𝑥 = ∫ 𝑥 𝑑𝑥 = 𝑥 . We then get
2 2
3

2 3 2 3
)2
∫ √𝑥 (ln 𝑥 𝑑𝑥 = ∫ (⏟
ln 𝑥 )2 (⏟√𝑥𝑑𝑥) = 𝑥 2 (ln 𝑥 ) − ∫ 𝑥 2 𝑑(ln 𝑥 )2
2

𝑢
3 3
𝑑𝑣

2 3 2
2 3 1
= 𝑥 (ln 𝑥 ) − ∫ 𝑥 (2(ln 𝑥 ) ( ) 𝑑𝑥)
2 2
3 3 𝑥
2 3 2
4 1
= 𝑥 2 (ln 𝑥 ) − ∫ 𝑥 2 ln 𝑥 𝑑𝑥
3 3
 Page 114 of 125

1
We use the integration by parts (with 𝑢(𝑥 ) = ln 𝑥, 𝑑𝑣 = 𝑥 𝑑𝑥2

1 3
2
⇒ 𝑣 = ∫ 𝑥 𝑑𝑥 = 𝑥 ) again to compute the second integral:
2 2
3

2 3 2
4 1 2 3 2
4 1
𝑥 2 (ln 𝑥 ) − ∫ 𝑥 2 ln 𝑥 𝑑𝑥 = 𝑥 2 (ln 𝑥 ) − ∫ (⏟ln 𝑥 ) (𝑥 2 𝑑𝑥)
3 3 3 3 ⏟
𝑢
𝑑𝑣

2 3 2
4 2 3 2 3
= 𝑥 2 (ln 𝑥 ) − [ 𝑥 2 ln 𝑥 − ∫ 𝑥 2 𝑑 (ln 𝑥 )]
3 3 3 3
2 3 2
8 3 8 1
= 𝑥 2 (ln 𝑥 ) − 𝑥 2 ln 𝑥 + ∫ 𝑥 2 𝑑𝑥
3 9 9
2 3 2
8 3 16 3
= 𝑥 2 (ln 𝑥 ) − 𝑥 2 ln 𝑥 + 𝑥 2 + 𝐶.
3 9 27
 Page 115 of 125

Review Example 3
Compute the integral

∫ 𝑥 cos2 (2𝑥 ) 𝑑𝑥

IDEA:
It may be better for us to express the term cos 2 𝑥 into sum of trigonometric
substitution using product-to-sum formula to get rid of the “square term”
before we use the integration by parts.

Solution
Using the product-to-sum formula, we have
cos2 2𝑥
=(cos 2𝑥)(cos 2𝑥)
1
∫ 𝑥 cos2 (2𝑥 ) 𝑑𝑥 =
⏞ ∫ 𝑥 [ (cos(2𝑥 + 2𝑥 ) + cos(2𝑥 − 2𝑥 ))] 𝑑𝑥
2
1 1 1 𝑥2
= ∫ 𝑥 cos 4𝑥 𝑑𝑥 + ∫ 𝑥𝑑𝑥 = ∫ 𝑥 cos 4𝑥 𝑑𝑥 + … . (∗)
2 2 2 4
 Page 116 of 125

It remains to compute the first integral. Using integration by parts with 𝑢(𝑥 ) = 𝑥
1
and 𝑑𝑣 = cos 4𝑥 𝑑𝑥 ⇒ 𝑣 = ∫ cos 4𝑥 𝑑𝑥 = sin 4𝑥, we have
4

1 1
∫⏟
𝑥⏟cos 4𝑥 𝑑𝑥 = ⏟
𝑥 ( sin 4𝑥) − ∫ sin 4𝑥 𝑑𝑥
⏟
⏟4 ⏟
4
𝑢 𝑑𝑣 𝑢 𝑑𝑢
𝑣 𝑣

∫ sin(𝑎𝑥+𝑏) 𝑑𝑥
1
=− cos(𝑎𝑥+𝑏)+𝐶
𝑎 1 1
=
⏞ 𝑥 sin 4𝑥 + cos 4𝑥 + 𝐶.
4 16
From (∗), we conclude that
2
1 1 𝑥
∫ 𝑥 cos2 (2𝑥 ) 𝑑𝑥 = 𝑥 sin 4𝑥 + cos 4𝑥 + + 𝐶.
8 32 4
 Page 117 of 125

Review Example 4
√𝑥 2
We let 𝑓 (𝑥 ) = ∫1 𝑒 −𝑡 𝑑𝑡.
(a) Find 𝑓′(𝑥).
(b) Hence, compute the integral
1
∫ √𝑥𝑓(𝑥 )𝑑𝑥 .
0
Solution of (a)
2
Let 𝐺 (𝑡) = ∫ 𝑒 −𝑡 𝑑𝑡, then by the fundamental theorem of calculus, we have

𝑓(𝑥 ) = 𝐺(√𝑥) − 𝐺 (1).

𝑑 𝑑 2 2
Using the fact that 𝐺 (𝑡) = ∫ 𝑒 −𝑡 𝑑𝑡 = 𝑒 −𝑡 , we have
𝑑𝑡 𝑑𝑡
take 𝑡=√𝑥
′( )
𝑑 𝑑 𝑑𝐺( √ 𝑥) 𝑑 √ 𝑥 −(√𝑥)
21
−
1
𝑓 𝑥 = 𝐺(√𝑥) − 𝐺 (1) = −0 =
⏞ 𝑒 𝑥 2
𝑑𝑥 𝑑𝑥 𝑑 √𝑥 𝑑𝑥 2
1 −1 −𝑥
= 𝑥 2𝑒 .
2
 Page 118 of 125

Solution of (b)
IDEA:
One cannot compute the integral directly since the expression of 𝑓(𝑥) is
unavailable. However since 𝑓′(𝑥) is known, one can use integration by
parts formula to transform 𝑓(𝑥) into 𝑓′(𝑥).

Using integration by parts with 𝑢(𝑥 ) = 𝑓(𝑥 ) and 𝑑𝑣 = √𝑥𝑑𝑥

3
2
⇒ 𝑣 = ∫ √𝑥𝑑𝑥 = 𝑥 , we have2
3
1 1 1
2 3 1
2 3
∫ √𝑥𝑓(𝑥 )𝑑𝑥 = ∫ 𝑓⏟
(𝑥 ) (⏟√𝑥𝑑𝑥) = 𝑥 2 𝑓(𝑥 )|0 − ∫ 𝑥 2 𝑑𝑓⏟ (𝑥 )
0 0
⏟
3 0 3
⏟
𝑢 𝑑𝑣 𝑑𝑢
𝑢𝑣|10 𝑣

2 2 2 1 3 ′
= (1) (1)
𝑓⏟ − (0)𝑓 (0) − ∫ 𝑥 2 𝑓 (𝑥 )𝑑𝑥
3 1 2
⏟
3 3 0
=∫1 𝑒 −𝑡 𝑑𝑡 =0 =0

2 1 3 1 −1 −𝑥
= − ∫ 𝑥 2 ( 𝑥 2 𝑒 ) 𝑑𝑥 (from (𝑎))
3 0 2
 Page 119 of 125

1 1 −𝑥 1 1
= − ∫ 𝑥𝑒 𝑑𝑥 = − ∫ ⏟𝑥 (⏟𝑒 −𝑥 𝑑𝑥 )
3 0 3 0 𝑢 𝑑𝑣
𝑢=𝑥
𝑑𝑣=𝑒 −𝑥 𝑑𝑥
⇒𝑣=∫ 𝑒 −𝑥 𝑑𝑥=−𝑒 −𝑥 1
1
=
⏞ − [(⏟−𝑥𝑒 −𝑥 )|10 − ∫ (⏟−𝑒 −𝑥 ) 𝑑𝑥
⏟]
3 1 0 𝑣 𝑑𝑢
𝑢𝑣|0

1
= − [−𝑒 −1 + (−𝑒 −𝑥 )|10 ]
3
2 −1 1
= 𝑒 − .
3 3
 Page 120 of 125

Review Examples on Integration Technique

Review Example 5
Compute the integral
tan−1 (𝑒 𝑥 )
∫ 𝑑𝑥
𝑒𝑥
Solution:
𝑑𝑦 1
We let 𝑦 = 𝑒 𝑥 . Then we have = 𝑒 𝑥 ⇒ 𝑑𝑥 = 𝑒 𝑥 𝑑𝑦. The integral becomes
𝑑𝑥
−1 ( 𝑥 )
tan 𝑒tan−1 (𝑒 𝑥 ) 1 tan−1 𝑦
∫ 𝑑𝑥 = ∫ ( 𝑥 𝑑𝑦) = ∫ 𝑑𝑦.
𝑒𝑥 𝑒𝑥 𝑒 𝑦2
Using integration by parts, we have
𝑑𝑣=𝑦 −2 𝑑𝑦
𝑣=∫ 𝑦 −2 𝑑𝑦=−𝑦 −1
tan−1 𝑦 −1 ( −2
∫ 𝑑𝑦 = ∫ tan
⏟ 𝑦 ⏟𝑦 𝑑𝑦) =
⏞ −𝑦 −1 tan−1 𝑦 − ∫ −𝑦 −1 𝑑 (tan−1 𝑦)
⏟
𝑦2 ⏟
𝑢 𝑑𝑣 𝑢𝑣
∫ 𝑣𝑑𝑢

−1 −1
1
= −𝑦 tan 𝑦+∫ 𝑑𝑦 … … (∗)
𝑦(1 + 𝑦 2 )
 Page 121 of 125

Since there is a quadratic factor in the denominator, we propose the following decomposition:
1 𝐴 𝐵𝑦 + 𝐶 2
= + ⇒ 1 = 𝐴 ( 𝑦 + 1) + 𝑦(𝐵𝑦 + 𝐶 ).
𝑦(1 + 𝑦 2 ) 𝑦 𝑦 2 + 1
Substitute 𝑦 = 0, we get 𝐴 = 1;
so 𝑦(𝐵𝑦 + 𝐶 ) = −𝑦 2 ⇒ 𝐵𝑦 + 𝐶 = −𝑦 ⇒ 𝐵 = −1, 𝐶 = 0.
Hence, we have
1 1 𝑦
∫ 𝑑𝑦 = ∫ 𝑑𝑦 − ∫ 𝑑𝑦
𝑦(1 + 𝑦 2 ) 𝑦 1 + 𝑦2
𝑑𝑢
𝑢=1+𝑦 2 ⇒ =2𝑦
𝑑𝑦
1 1 1 1
=
⏞ ln|𝑦| − ∫ 𝑑𝑢 = ln|𝑦| − ln|𝑢| + 𝐶 = ln|𝑦| − ln|1 + 𝑦 2 | + 𝐶
2 𝑢 2 2
Hence from (*), we conclude that
tan−1 (𝑒 𝑥 ) tan−1 𝑦 −1 −1
1
∫ 𝑥
𝑑𝑥 = ∫ 2
𝑑𝑦 = −𝑦 tan 𝑦 + ln |𝑦 | − ln|1 + 𝑦 2 | + 𝐶
𝑒 𝑦 2
tan−1 (𝑒 𝑥 ) 𝑥|
1 2𝑥 |
tan −1 ( 𝑥 )
𝑒 1
=− 𝑥
+ ln|𝑒 − ln|1 + 𝑒 + 𝐶 = − 𝑥
+ 𝑥 − ln|1 + 𝑒 2𝑥 | + 𝐶.
𝑒 2 𝑒 2
 Page 122 of 125

Review Example 6
Compute the integral
1
𝑥+4
∫ 𝑑𝑥
0 √𝑥 2 + 2𝑥 + 2
Solution:
𝑑𝑦 1
We let 𝑦 = 𝑥 2 + 2𝑥 + 2, then = 2𝑥 + 2 ⇒ 𝑑𝑥 = 2𝑥+2 𝑑𝑦.
𝑑𝑥

When 𝑥 = 1, 𝑦 = 12 + 2(1) + 2 = 5; When 𝑥 = 0, 𝑦 = 2.

Then we split the integral as
1 1 1
𝑥+4 𝑥+1 3
∫ 𝑑𝑥 = ∫ 𝑑𝑥 + ∫ 𝑑𝑥 … (∗)
0 √𝑥 2 + 2𝑥 + 2 0 √𝑥 2 + 2𝑥 + 2 0 √𝑥 2 + 2𝑥 + 2
For the first integral, we use the proposed substitution and get
1 5
𝑥+1 𝑥+11 1 5 1 1 5
∫ 𝑑𝑥 = ∫ ( 𝑑𝑦) = ∫ 𝑑𝑦 = [𝑦 2 ] = √5 − √2.
2
0 √𝑥 + 2𝑥 + 2
2
2 √𝑥 + 2𝑥 + 2 2𝑥 + 2 2 2 √𝑦 2
 Page 123 of 125

For the second integral, we need to propose another substitution (probably trigonometric
substitution) since the original substitution does not work. The second integral becomes
1 1
3 1
∫ 𝑑𝑥 = 3 ∫ 𝑑𝑥
0 √𝑥 2 + 2𝑥 + 2 0 √(𝑥 + 1)2 +1
𝑑𝑥
We let 𝑥 + 1 = tan 𝜃 ⇒ 𝑥 = tan 𝜃 − 1. Then = sec 2 𝜃.
𝑑𝜃
𝜋
When 𝑥 = 1, 𝜃 = tan−1 2; When 𝑥 = 0, 𝜃 = tan−1 1 = 4 . Then

1 tan−1 2 tan−1 2
1 1
3∫ 𝑑𝑥 = 3 ∫ sec 2 𝜃 𝑑𝜃 = 3 ∫ sec 𝜃 𝑑𝜃
𝜋 √tan2 𝜃 +1 𝜋
0 √(𝑥 + 1)2 +1 4 4
−1 2 𝜋 𝜋
= 3 ln|sec 𝜃 + tan 𝜃| |tan
𝜋 = 3 ln|sec(tan−1 2) + tan(tan−1 2)| − 3 ln |sec ( ) + tan ( )|
4 4 4
sec 𝜃=√1+tan2 𝜃
=
⏞ 3 ln|√5 + 2| − 3 ln|√2 + 1|.

Thus, we can conclude from (*) that

1
𝑥+4
∫ 𝑑𝑥 = √5 − √2 + 3 ln|√5 + 2| − 3 ln|√2 + 1|.
0 √𝑥 2 + 2𝑥 + 2
 Page 124 of 125

Review Example 7
Compute the integral
ln(𝑥 2 + 2𝑥 + 10)
∫ 𝑑𝑥.
(2𝑥 + 1)2
Solution: Using integration by parts, we have
ln(𝑥 2 + 2𝑥 + 10)
∫ ln(𝑥 2 + 2𝑥 + 10) [(
𝑑𝑥 = ∫ ⏟ ⏟2𝑥 + 1)−2 𝑑𝑥 ]
(2𝑥 + 1) 2
𝑢 𝑑𝑣
1
𝑣=−2(2𝑥+1)−1
1 −1 ( 2
1
=
⏞ (
− 2𝑥 + 1 ) ln 𝑥 + 2𝑥 + 10 − − (2𝑥 + 1)−1 𝑑 ln(𝑥 2 + 2𝑥 + 10)
) ∫
2 2
1 1 2𝑥 + 2
= − (2𝑥 + 1)−1 ln(𝑥 2 + 2𝑥 + 10) + ∫ 𝑑𝑥 … (∗)
2 2 (2𝑥 + 1)(𝑥 2 + 2𝑥 + 10)
Since the integrand is a proper rational function and the denominator contains a quadratic
factor, we shall propose the following decomposition:
2𝑥 + 2 𝐴 𝐵𝑥 + 𝐶
= + .
(2𝑥 + 1)(𝑥 2 + 2𝑥 + 10) 2𝑥 + 1 𝑥 2 + 2𝑥 + 10
 Page 125 of 125

⇒ 2𝑥 + 2 = 𝐴(𝑥 2 + 2𝑥 + 10) + (2𝑥 + 1)(𝐵𝑥 + 𝐶)

1 37 4
Substitute 𝑥 = − 2, we get 1 = 𝐴 ( 4 ) ⇒ 𝐴 = 37.
2
Compare the coefficient of 𝑥 2 : 0 = 𝐴 + 2𝐵 ⇒ 𝐵 = − 37 ;
34
Compare the constant term: 2 = 10𝐴 + 𝐶 ⇒ 𝐶 = 37.
Hence, we can decompose the integral as
2𝑥 + 2 4 1 1 −2𝑥 + 34
∫ 𝑑𝑥 = ∫ 𝑑𝑥 + ∫ 𝑑𝑥
(2𝑥 + 1)(𝑥 2 + 2𝑥 + 10) 37 2𝑥 + 1 37 𝑥 2 + 2𝑥 + 10
4 1 1 2𝑥 + 2 1 36
= ∫ 𝑑𝑥 − ∫ 2 𝑑𝑥 + ∫ 2 𝑑𝑥.
37 2𝑥 + 1 37 𝑥 + 2𝑥 + 10 37 𝑥 + 2𝑥 + 10
𝑦=𝑥 2 +2𝑥+10
4 1 1 1 36 1
=
⏞ ( ln|2𝑥 + 1|) − ∫ 𝑑𝑦 + ∫ 𝑑𝑥
37 2 37 𝑦 37 (𝑥 + 1)2 + 9
2 1 4 1
= ln|2𝑥 + 1| − ln|𝑦| + ∫ 𝑑𝑥
37 37 37 𝑥+1 2
( 3 ) +1

2 1 2
12 −1
𝑥+1
= ln|2𝑥 + 1| − ln|𝑥 + 2𝑥 + 10| + tan ( )+𝐶
37 37 37 3