1
HKDSE Mathematics
Indefinite Integration
By Leon Lee
1. Introduction
Integration, either definite or indefinite, is a large topic in HKDSE Extended Module 2. With the
introduction of integrations by substitution and by parts, which were not required in HKCEE
Additional Mathematics, to the syllabus, a much wider variety of questions can be set compared
with the old syllabus. Therefore, students should pay much more attention to this topic, and try to
master this topic as well as possible for good results in M2 examination.
This note will demonstrate the techniques in solving problems involving indefinite integration as
detailed as possible. It will start from the basic problems, and gradually to the hardest problems
which involve advanced techniques in integration.
2. What is Indefinite Integration?
→ Indefinite integration can be considered the ‘reverse’ process of differentiation.
→ In differentiation, we find the derivative of a function.
In indefinite integration, we find the primitive function of a function.
→ In simpler words, if the derivative of )(xF is )(xf , then )(xF is a primitive function of )(xf ,
i.e. )()( xfxFdx
d= .
→ The primitive function of a function is not unique in nature.
Note that )(]2)([]1)([)( xfxFdx
dxF
dx
dxF
dx
d=+=+= .
Then )(xF , 1)( +xF and 2)( +xF are all primitive functions of )(xf .
When we replace 1 or 2 by any other real constants, sayπ , e , 2 or 10− , we still get the
same result.
→ From the above results, we see that there are infinitely many primitive functions for any
integrable function.
For convenience, if )()( xfxFdx
d= , then we write ∫ += CxFdxxf )()( .
→ The constant C is called the constant of integration, and it is
arbitrary in nature.
→ The sign ∫ is called the integral sign, and )(xf is called
the integrand.
Quick Example:
As x
x
dx
d
4
1
2=
, we have C
xdx
x+=∫
24
1.
F(x)+C f (x)
differentiation
indefinite integration
Figure 1 A diagrammatic representation of
differentiation and indefinite integration
2
3. Methods of Indefinite Integration
The methods of indefinite integration will be introduced below.
3.1. Elementary Integration
The following shows some fundamental indefinite integration results, which can be obtained
directly from differentiation.
(1) constant a is where, kCkxkdx∫ +=
(2) ∫ −≠+
+
=
+
1 where,1
1
nCn
xdxx
n
n
(3) ∫ ≠+= 0 where,ln1
xCxdxx
(4) 0 where, ≠+=∫ aCa
edxe
ax
ax
From (4), when 1=a , we have
(5) ∫ += Cedxexx .
The following gives the proof of (3). The remaining is left to readers as an exercise.
Proof:
When 0>x , xx lnln = . We havex
xdx
d 1)(ln = .
When 0<x , )ln(ln xx −= . We havexx
xdx
d 1)1(
1)][ln( =−
−
=− .
Combining the results, we have Cxdxx
+=∫ ln1
.
Furthermore, we have the following properties. They can also be easily proved by employing
differentiation.
(6) ∫ ∫= constant a is where,)()( kdxxfkdxxkf
(7) [ ]∫ ∫ ∫±=± dxxgdxxfdxxgxf )()()()(
Using (7) repeatedly, we have
(8) [ ] ∫∫∫ ∫ ±±±=±±± dxxfdxxfdxxfdxxfxfxfnn
)(...)()()(...)()(2121
Important!
is the absolute sign. It is defined as:
<−
≥=
0 when ,
0 when ,
xx
xx
x
3
Example 1
Find dxx
x∫
−
4
2 1. (HKDSE Sample Paper)
Solution:
Cx
xxxx
Cx
xxxx
dxxx
xxx
dxxx
xxx
dxxx
xCx
xCx
xCxdxx
x
+−−+−=
+−
+−⋅+⋅−=
+−+−=
+−+−=
−+
−+
−+
−+=
−
−
−
∫
∫
∫∫
3
3
69
3369
4258
4
258
43
24
3
2
224
2
324
1
42
4
2
3
1ln42
3
2
9
3ln4
36
64
9
)4
64(
1464
11)(
1)(
1)()(
1
Note: Always add the constant C after indefinite integration!
To make it simple, you should add it whenever the integral sign ∫ disappears.
Example 2
Find (a) ( ) dxx∫ +
2
1 (b) ∫ dxx10 .
Solution:
(a) ( )∫ + dxx2
1
Cxxx
Cxxx
dxxx
dxxx
+++=
++⋅+=
++=
++=
∫
∫
2
32
2
32
2
1
3
4
2
3
22
2
)12(
)12(
(b) We first let constant. a is where,10 aeaxx =
Solving, we have 10ln=a .
C
Ce
dxe
dx
x
x
x
x
+=
+=
= ∫
∫
10ln
10
10ln
10
10ln
10ln
10ln
10ln
ln10ln
=
=
=
a
axx
eaxx
(Binomial Theorem)
4
3.2. Integration by Substitution
The above formulae are very limited in usage and cannot deal with most integrals like
dxx
x
∫−
+
1
1 or dx
x
x
∫+1
2
2
. We need a new method called ‘integration by substitution’ to deal
with these integrals.
Let )( xgu = . Then, we have ∫ ∫= duufdxxgxgf )()('))(( .
The proof is given in the appendix of this note on p.46.
This method can be regarded as the ‘reverse’ of the chain rule in differentiation.
Refer to the following illustration to see how we can apply this method to find integrals:
Consider dx
x
x
∫+1
2
2
. We take 1)( 2+== xxgu . Then, we have xxg' 2)( = .
Cx
Cu
duu
dxxgxg
xdxxg
dxx
x
++=
+=
=
=
⋅=
+
∫
∫
∫∫
−
12
2
)(')(
1
2)(
1
1
2
2
2
1
2
1
2
As seen from the above example, we can see that we express the integral in the form
∫ dxxgxgf )('))(( and transform it into the form ∫ duuf )( for simpler calculation.
However, we usually use the following way:
Let 12+= xu , then xdxdu 2= , i.e. du
xdx
2
1= .
Cx
Cu
duu
duxu
xdx
x
x
++=
+=
=
⋅=
+
∫
∫∫
−
12
2
2
12
1
2
2
2
1
2
1
2
We use a suitable substitution )(xgu = at the beginning. We differentiate it with respect to x
(w.r.t. x, in short) to obtain dxxgdu )('= . By rearranging terms, we get duxg
dx)('
1= . We put
it into the original integral and express the whole integral in terms of the new variable u only.
Finally, we find the integral, and express the result in terms of the original variable x.
We express dx in
terms of du.
Here, x
xf1
)( = and 1)( 2+= xxg .
5
Example 3 (2012 DSE)
(a) Find ∫+
dxx
x 1.
(b) Using the substitution 12−= xu , find ∫
−
dxx
x
12
3
.
Solution:
(a) ∫∫ ++=+=+
Cxxdxx
dxx
xln)
11(
1
(b) Let 12−= xu , then xdxdu 2= , i.e. du
xdx
2
1= .
2
1 where,'1ln
2
1
2
1
1ln2
1)1(
2
1
ln2
1
2
1
1
2
1
1
22
22
2
3
−=+−+=
+−+−=
++=
+=
−∫∫
CC'Cxx
Cxx
Cuu
duu
udx
x
x
*Think About*:
(1) Why did we not to include 2
1− in final answer and use a new constant 'C instead?
(2) Why and how is 2x changed into 1+u ?
Example 4
Suppose 0>x .
Find (a) ∫+
dxebax (b) ∫
++
+ dxebaxcbxax
2
)2( (c) ∫ dxex
x
2
2
1 (d) ∫ dx
x
xln.
Solution:
(a) ∫∫ +=+=
+
++
Ca
ebaxde
adxe
bax
baxbax)(
1
(b) Cecbxaxdedxebaxcbxaxcbxaxcbxax+=++=+
++++++
∫∫222
)()2(2
(c) Cex
dedxex
xxx +−=
−= ∫∫
222
22
12
2
11
(d) Cx
xxddxx
x+== ∫∫ 2
)(ln)(lnln
ln 2
(by (a)) Always remember:
Express your final answer
in terms of the original
variable!
6
Example 5
Find ∫ + dxxx20132 )1( .
Solution:
Let 1+= xu , then dxdu = .
Cxxx
Cuuu
duuuu
duuuu
duuu
dxxx
++
++
−+
=
++−=
+−=
+−=
−=
+
∫
∫
∫
∫
2014
)1(
2015
)1(2
2016
)1(
20142015
2
2016
)2(
)12(
)1(
)1(
201420152016
201420152016
201320142015
20132
20132
20132
Example 6
Find (a) ∫−
+dx
x
x
1
1 (b) ∫
−
dxx
x
1
2
.
Solution:
(a) ∫∫−
+−=
−
+dx
x
xdx
x
x
1
21
1
1
Cxx
xdx
dx
dxxx
x
+−+=
−−
+=
−+
−
−=
∫ ∫
∫
1ln2
)1(1
12
1
2
1
1
(b) Let 1−= xu , then dxdu = .
∫−
dxx
x
1
2
'1ln22
)1(
1ln)1(22
)1(
ln22
)1
2(
12
)1(
2
2
2
2
2
Cxxx
Cxxx
Cuuu
duu
u
duu
uu
duu
u
+−++−
=
+−+−+−
=
+++=
++=
++=
+=
∫
∫
∫
We can write ∫ dx1 simply as
∫ dx , omitting the constant 1.
7
Example 7 (Long Division & Partial Fractions)
(a) Let 23
)()(
23
18622
23456
+−
+≡
+−
++−−−−
xx
xgxf
xx
xxxxxx, where )(xf and )(xg are
two polynomials with 4)(deg ≤xf and 2)(deg <xg . Find )(xf and )(xg .
(b) Let 2123
)(2
−
+
−
≡
+− x
B
x
A
xx
xg, where A and B are constants. Find A and B.
(c) Hence, find ∫+−
++−−−−dx
xx
xxxxxx
23
186
2
23456
.
Solution:
(a) Here, we perform long division.
43 2 1
1
8124
1 8 124
6 93
1 8 6 53
462
1 8 6 1 3 2
23 1
1 8 6 1 1 1 12 3 1
+++
+
+−+
++−+
+−+
++−−+
+−+
++−−−+
+−
++−−−−+−
Then, we have 23
1432
23
186
2
234
2
23456
+−
++++=
+−
++−−−−
xx
xxxx
xx
xxxxxx
.
Thus, xxxxxf 432)( 234+++≡ and 1)( ≡xg .
(b) From (a), 1)( ≡xg .
)1()2(1
2123
12
−+−≡
−
+
−
≡
+−
xBxA
x
B
x
A
xx
Put 1=x , we have )11()21(1 −+−= BA , i.e. 1−=A .
Put 2=x , we have )12()22(1 −+−= BA , i.e. 1=B .
(c) From above, we have1
1
2
1432
23
186 234
2
23456
−
−
−
++++=
+−
++−−−−
xx
xxxx
xx
xxxxxx
Cx
xxx
xx
Cxxxxxx
dxxx
xxxxdxxx
xxxxxx
+−
−++++=
+−−−++++=
−−
−++++=
+−
++−−−−∫∫
1
2ln2
25
1ln2ln225
1
1
2
1432
23
186
23
45
23
45
234
2
23456
Points to note:
1. The notation ‘deg’ is used to refer to the
degree of a polynomial. For instance, let
12)( 25+−= xxxh , then )(deg xh is
equal to 5.
2. This type of long division without
variables written is known as the method of
detached coefficients.
3. The method employed in (b), i.e. to break
an algebraic fraction into sum of fractions
with denominators of smaller degrees, is
called resolving an algebraic fraction into
partial fractions.
8
Example 8 (Reduction Formula)
Let nxxy )(ln= , where 0>x .
(a) Find dx
dy.
(b) Let ∫= dxxIn
n)(ln , where 0≥n .
Using the result in (a), or otherwise, show that 1
)(ln−
−=n
n
nnIxxI
(c) Hence, find 4I .
Solution:
(a) nnnn xxnxx
xnxdx
dy)(ln)(ln)1()(ln
1)(ln 11
+=⋅+⋅⋅=−−
(b) Integrating both sides w.r.t. x, we have
1
1
1
1
)(ln
)(ln)(ln)(ln
)(ln)(ln)(ln
)(ln)(ln
−
−
−
−
−=
−=
+=
+=
∫∫
∫∫
∫∫
n
n
n
nnn
nnn
nn
nIxxI
dxxnxxdxx
dxxdxxnxx
dxxdxxny
(c) Using the result in (b) repeatedly, we have
[ ][ ]
[ ]
Cxxxxxxxxx
dxxxxxxxxxx
Ixxxxxxxx
Ixxxxxx
Ixxxx
IxxI
++−+−=
+−+−=
−−+−=
−+−=
−−=
−=
∫24ln24)(ln12)(ln4)(ln
)(ln24ln24)(ln12)(ln4)(ln
ln24)(ln12)(ln4)(ln
2)(ln12)(ln4)(ln
3)(ln4)(ln
4)(ln
234
0234
0
234
1
234
2
34
3
4
4
Notes:
(1) The method employed in this question is to establish a reduction formula, which
reduces an integral from a higher to lower power, with the same form. A reduction
formula can be used repeatedly to lower the power of the integral, until it can be
integrated easily. In this way, the integrals in certain forms, no matter how high the
power is, can be found.
(2) We can find out a reduction formula by differentiation. However, we usually use a
technique called integration by parts to find them out. This will be introduced in
Section 3.5. (Refer to Example 34 for details)
9
3.3. Integration of Trigonometric Functions
Trigonometric functions can be differentiated. Similarly, they can also be integrated.
3.3.1. Basic Integration of Trigonometric Functions
From differentiation, we have the following basic results.
(1) ∫ +−= Cxxdx cossin
(3) ∫ += Cxxdx tansec2
(5) ∫ += Cxxdxx sectansec
For other integrals of trigonometric functions, different trigonometric identities are useful.
Example 9
Find (a) ∫ dxx
2sin
2 (b) ∫ xdx2tan .
Solution:
(a) Cxxxdxxxdxdxdxx
dxx
+−=−=−=−
= ∫∫∫∫∫ sin2
1
2
1cos
2
1
2
1cos
2
1
2
1
2
cos1
2sin
2
(b) Cxxdxxxdx +−=−= ∫∫ tan)1(sectan 22
(Can you find ∫ dxx
2cos
2 and ∫ xdx2cot ? Give them a try!)
3.3.2. Integration of Trigonometric Functions with Method of Substitution
The method of substitution can also be used in finding integrals of trigonometric functions.
Example 10
Find ∫ xdxx cos3sin .
Solution:
[ ]
Cxx
xxdxxd
dxxx
dxxxxx
xdxx
+−−=
⋅+⋅=
+=
−++=
∫∫
∫
∫
∫
2cos4
14cos
8
1
)2(2sin2
1
2
1)4(4sin
4
1
2
1
)2sin4(sin2
1
)3sin()3sin(2
1
cos3sin
(2) ∫ += Cxxdx sincos
(4) ∫ +−= Cxxdx cotcsc2
(6) ∫ +−= Cxxdxx csccotcsc
Point to note:
Generally, we have the following formulas
for constants a and b, with 0≠a :
Cbaxa
dxbax
Cbaxa
dxbax
++=+
++−=+
∫
∫
)sin(1
)cos(
)cos(1
)sin(
They can be proven easily by differentiation
or method of substitution, with substitution
baxu += . They are widely accepted
and are not required to be proven again
in your calculations.
10
Example 11
(a) Prove that x
x
x
2sin
2cos1tan
−
= .
(b) Hence, or otherwise, find ∫−
dxx
x
2sin
2cos1.
Solution:
(a) LHStancos
sin
cossin
sin
2
2cos1
cossin2
2
2sin
2cos1RHS
2
====
−
⋅=
−
= x
x
x
xx
xx
xxx
x
Thus, we have x
x
x
2sin
2cos1tan
−
= .
(b) Cxxdx
dxx
xxdxdx
x
x+−=−===
−
∫ ∫∫∫ cosln)(coscos
1
cos
sintan
2sin
2cos1
Example 12
Find ∫ xdxsec .
Solution:
∫ xdxsec
∫+
+= dx
xx
xxx
tansec
)tan(secsec
∫+
+
=
xx
xxd
tansec
)tan(sec (Here, xxu tansec += and dxxxxdu )tan(secsec += .)
Cxx ++= tansecln
Notes:
(1) Students are strongly recommended to recite the substitution used in Example 12, as well
as the way to find this integral.
(2) Alternatively, you can find the integral using partial fractions as follows:
∫ ∫∫∫∫−
=
−
=== duu
xdx
dxx
xdx
xxdx
222 1
1)(sin
sin1
1
cos
cos
cos
1sec (for xu sin= )
We can let u
B
u
A
u −
+
+
≡
− 111
1
2. Solving the simultaneous equations set up by
comparing like terms, we can get 2
1=A and
2
1=B . (Refer to Example 7 for details)
Then, Cx
xCuudu
udu
uxdx +
−
+=+−−+=
−
+
+
= ∫∫∫sin1
sin1ln
2
11ln
2
11ln
2
1
1
1
2
1
1
1
2
1sec .
(3) Can you find ∫ xdxcsc ? Give it a try! (Hint: Let xxu cotcsc += )
(or Cx +secln )
11
Example 13 (Subsidiary Angle)
(a) Let2
0π
θ << . By differentiating )tanln(sec θθ + , show that ∫ ++= Cd )tanln(secsec θθθθ ,
where C is a constant.
(b) Let )cos(sincos αθθθ +=− r , where r and α are constants with 0>r and 22
π
α
π
<<− .
Find r and α .
(c) Hence, find ∫−
θθθd
sincos
1.
Solution:
(a) θθθθθθ
θθθ
sec)sectan(sectansec
1)tanln(sec 2
=+
+
=+
d
d
Integrating both sides w.r.t. θ , we have
Cd ++=∫ )tanln(secsec θθθθ
(b) θααθαθθθ sincossincos)cos(sincos rrr −=+=−
Comparing like terms, we have
−=−
=
)2...(....................1cos
..(1).................... 1sin
α
α
r
r
1tan:)2(
)1(=α
We get 4
π
α = .
2222222 11)cos(sin:)2()1( +=++ ααr
We get 2=r .
(c) From (b), we have
+=−4
cos2sincosπ
θθθ .
C
d
dd
+
++
+=
+
+=
+
=−
∫
∫∫
4tan
4secln
2
2
44sec
2
1
4cos
1
2
1
sincos
1
πθ
πθ
πθ
πθ
θπ
θ
θθθ
Notes:
(1) The form )sin( α±xr or )cos( α±xr is called the subsidiary angle form. It expresses the sum
(or difference) of two sine (or cosine) functions (with any non-zero coefficients) as a product
of a non-zero real number and a sine (or cosine) function.
(2) Subsidiary angle form is NOT required in HKDSE M2 Examination. However, question
format as in Example 13 is possible to appear in the examination.
12
Example 14
Find (a) ∫ dxx
x)sin(ln (b) ∫ dxxx )(cossecsin 2 (c) ∫ ++ dxxxx )1cot()1csc( 22 .
Solution:
(a) ∫ dxx
x)sin(ln
Cx
xdx
+−=
= ∫)cos(ln
)(ln)sin(ln
(b) ∫ dxxx )(cossecsin 2
Cx
xdx
+−=
−= ∫)tan(cos
)(cos)(cossec2
(c) ∫ ++ dxxxx )1cot()1csc( 22
Cx
xdxx
++−=
+++= ∫
)1csc(2
1
)1()1cot()1csc(2
1
2
222
3.3.3. Integration of Special Types of Trigonometric Functions
The integrals of some trigonometric functions can be found easily by certain methods.
(Here, p and q are positive integers.)
Type A: ∫ xdxxqp
cossin
Example 15
Find (a) ∫ xdxx cossin 5 (b) ∫ xdxx36 cossin (c) ∫ xdxx
46 cossin .
Solution:
(a) Cx
xxdxdxx +== ∫∫ 6
sin)(sinsincossin
6
55
(b) ∫ xdxx36 cossin
Cxx
xxdxxd
xdxx
xdxxx
+−=
−=
−=
⋅⋅=
∫ ∫
∫
∫
9
sin
7
sin
)(sinsin)(sinsin
)(sin)sin1(sin
coscossin
97
86
26
26
13
(c) We first express xx46
cossin as a sum of nxcos , where n is an integer.
( )
)10cos8cos26cos34cos82cos26(512
1
6cos2
110cos
2
18cos6cos22cos24cos42cos33
256
1
)8cos2cos8cos4cos2cos44cos42cos33(256
1
)2cos1(8cos4cos43256
1
)2cos1(2
8cos14cos21
128
1
)2cos1)(4cos4cos21(128
1
)2cos1(2
4cos1
32
1
)2cos1(2sin32
1
2
2cos12sin
2
1
sin)cos(sincossin
2
2
4
4
2446
xxxxx
xxxxxxx
xxxxxxx
xxx
x
x
x
xxx
x
x
xx
x
x
xxxxx
−++−−=
−−+++−−=
−++−−=
−+−=
−
++−=
−+−=
−
−=
−=
−
=
=
Cxxxxxx
Cxxxxxx
dxxxxxxxdxx
+−++−−=
+
−++−−=
−++−−= ∫∫
10sin5120
18sin
2048
16sin
1024
14sin
256
12sin
512
1
256
3
10sin10
18sin
4
16sin
2
14sin22sin6
512
1
)10cos8cos26cos34cos82cos26(512
1cossin 46
Notes:
(1) If at least one of p or q is odd, we use the substitution xu sin= or xu cos= , in order to
convert the integral into a form only with the substitution used. In Examples 13(a) and 13(b),
we used xu sin= as the substitution and the whole integral was expressed in terms of xsin .
(2) If both p and q are even, we use the identities2
2cos1sin
2 x
x
−
= and2
2cos1cos
2 x
x
+=
repeatedly to lower the power of the integral. The integral is eventually expressed as a sum of
sine and cosine functions, and can then be integrated easily.
14
Type B: ∫ xdxxqp
sectan
Example 16
Find (a) ∫ xdxx25 sectan (b) ∫ xdxx
42 sectan (c) ∫ xdxx53 sectan .
Solution:
(a) ∫ xdxx25 sectan
Cx
xxd
+=
= ∫
6
tan
)(tantan
6
5
(b) ∫ xdxx42 sectan
Cxx
xxdxxd
xdxx
xdxxx
++=
+=
+=
⋅=
∫∫
∫
∫
3
tan
5
tan
)(tantan)(tantan
)(tan)1(tantan
secsectan
35
24
22
222
(c) ∫ xdxx53 sectan
Cxx
xxdxxd
xxdx
xdxxxx
+−=
−=
−=
⋅=
∫∫
∫
∫
5
sec
7
sec
)(secsec)(secsec
)(secsec)1(sec
tansecsectan
57
46
42
42
Notes:
(1) If q is even, we use the substitution xu tan= and convert the whole integral in terms of xtan .
Besides this, xn
sec (where n is an even number) can be expressed in terms of xtan using the
relation xx22
tan1sec += .
(2) If both p and q are odd, we use the substitution xu sec= and convert the whole integral in
terms of xsec . Besides this, xn
tan (where n is an even number) can be expressed in terms of
xsec using the relation 1sectan22
−= xx .
(3) It is much more difficult to find the integral if p is even and q is odd. A new technique called
‘integration by parts’ is required, and it will be introduced in Section 3.5.
(Refer to Example 33 on p. 33 for details.)
(4) Integrals in the form ∫ xdxxqp
csccot are found in similar way.
15
Type C: ∫ xdxn
sin , ∫ xdxn
cos , etc.
The techniques employed are similar to those used in finding Type A and Type B integrals.
Example 17
(a) Find (i) ∫ xdx4
sin (ii) ∫ xdx5cos (iii) ∫ xdx
4tan (iv) ∫ xdx5
cot .
(b) By differentiating xxn
cossin1− , show that ∫∫
−−−
+−= xdxn
nxx
nxdx
nnn 21sin
1cossin
1sin ,
where 2≥n . Hence, find ∫ xdx4
sin .
Solution:
(a) (i) ∫ xdx4
sin
1
2
2
4sin32
12sin
4
1
8
3
)4cos2cos43(8
1
2
4cos12cos21
4
1
)2cos2cos21(4
1
2
2cos1
Cxxx
dxxx
dxx
x
dxxx
dxx
++−=
+−=
++−=
+−=
−=
∫
∫
∫
∫
(ii) ∫ xdx5cos
Cxxx
xdxx
xdxx
++−=
+−=
−=
∫
∫
53
42
22
sin5
1sin
3
2sin
)(sin)sinsin21(
cos)sin1(
(iii) ∫ xdx4tan
Cxxx
dxxxxd
xdxxdxx
xdxx
++−=
−−=
−=
−=
∫∫
∫∫
∫
tan3
tan
)1(sec)(tantan
tantansec
tan)1(sec
3
22
222
22
We try to express the integral
as a sum of cosine functions.
We try to express the integral
in terms of xsin .
16
(iv) ∫ xdx5
cot
Cxxx
xdx
xx
dxx
xxxdxxd
xdxxdxxxdxx
xdxx
+++−=
++−=
++−=
+−=
⋅−=
∫
∫∫∫
∫∫∫
∫
sinlncsc4
csc
)(sinsin
1csc
4
csc
sin
cos)(csccsc2)(csccsc
cotcotcsc2cotcsc
cot)1(csc
2
4
2
4
3
24
22
(b) Let xxyn
cossin1−
= .
xnxn
xxnx
xxnx
xxnxxxdx
dy
nn
nn
nn
nn
sinsin)1(
)sin1(sin)1(sin
cossin)1(sin
))(cos)(sin1)((cos)sin(sin
2
22
22
21
−−=
−−+−=
−+−=
−+−=
−
−
−
−−
Integrating both sides w.r.t. x, we have
∫∫
∫∫
∫∫
−−
−−
−
−+−=
−−=
−−=
xdxn
nxx
nxdx
xxdxnxdxn
xdxnxdxny
nnn
nnn
nn
21
12
2
sin1
cossin1
sin
cossinsin)1(sin
sinsin)1(
2
3
03
234
8
3cossin
8
3cossin
4
1
)(sin2
1cossin
2
1
4
3cossin
4
1
sin4
3cossin
4
1sin
Cxxxxx
dxxxxxx
xdxxxxdx
++−−=
+−+−=
+−=
∫
∫∫
An integral may take up different forms. We obtain different results for the same integral ∫ xdx4
sin .
If we subtract the result in (a)(i) from that in (b), we get:
0
2cos2sin16
12sin
4
12sin
16
32cos2sin
16
12sin
16
1
2cos2sin16
12sin
4
12sin
16
3
2
2cos12sin
2
1
4
1
4sin32
12sin
4
1
8
3
8
3cossin
8
3cossin
4
1 3
=
−+−+−=
−+−
−
−=
+−−
+−−
xxxxxxx
xxxx
x
x
xxxxxxxx
It follows that C1 = C2. There is no difference between the two integrals, and they are solely equivalent.
17
3.4. Integration by Trigonometric Substitution
Integration of trigonometric functions can be helpful in finding certain indefinite integrals. Before
learning how to do so, we have to learn about inverse trigonometric functions.
3.4.1. Inverse Trigonometric Functions
We have come across with a lot of trigonometric functions, say, xy sin= . To express x in terms of y,
we define yx1
sin−
= (or yx arcsin= ), where y1sin − is read as ‘arcsine y’.
(a) Principal Values
Consider xy sin= . For2
2=y , we may have ,...
4
9,
4
3,
4
πππ
=x (in fact, 4
)1(π
πn
nx −+= ,
where n is an integer). A function should give one and only one output for every input (e.g.
for yx1
sin−
= , if x is a function of y, then for any value of y, there is one and only one
corresponding value of x). To make sure the relationship, we define an interval called
‘principal values’ to restrict the value of x.
Refer to the following illustration:
Let xy1
sin−
= . (Note that x and y is exchanged here, unlike the above example)
For arcsine functions, the principal values are
−
2,
2
ππ
(i.e. 22
ππ
≤≤− x ).
When 2
1=x , we have
6
π
=y , i.e. 62
1sin
1 π
=− .
When 1−=x , we have 2
π
−=y , i.e. 2
)1(sin 1 π
−=−− .
Note: Our calculators are user-friendly. When you enter inverse trigonometric functions, it will
automatically show the principal value as the answer.
The principal values of different functions are shown in the table below:
Function Principal Values
x1
sin−
−
2,
2
ππ
x1
cos− [ ]π,0
x1
tan−
−
2,
2
ππ
x1
csc−
−−2
,π
π or
2,0π
x1
sec−
−−2
,π
π or
2,0π
x1
cot− ),0( π
Note: 02
cot =π
.
Point to note:
[ ] refers to closed interval (i.e. including the
limits) while ( ) refers to open interval (i.e.
excluding the limits).
Examples:
(1)
−∈
2,
2
ππ
x means 22
ππ
≤≤− x .
(2)
−∈
2,
2
ππ
x means 22
ππ
<<− x .
(3)
−∈
2,
2
ππ
x means 22
ππ
≤<− x .
18
Example 18
Find (a) 75.0sin1− (b)
4cos
1 π− (c) 3.3cot1− , in radian measures, correct to 3 decimal places.
Solution:
(a) 848.075.0sin1
≈−
(i.e. 75.0848.0sin ≈ )
(b) 667.04
cos1
≈−π
(i.e. 4
667.0cosπ
≈ )
(c) We cannot enter ‘ 1cot
− ’ directly into our calculators.
Therefore, we use the following method.
294.0
3.3
1tan3.3cot
11
≈
= −−
(i.e. 3.3294.0cot ≈ )
(b) Graphs of Inverse Trigonometric Functions
The following shows the graphs of the six trigonometric functions and those of their inverse
functions for comparison:
Let 3.3cot1−
=x .
Then, we have
3.3cot =x
3.3
1tan
3.3
1tan
3.3tan
1
1−=
=
=
x
x
x
i.e.
= −−
3.3
1tan3.3cot
11
x
y
O
y=sin x
y=sin−1
x
1 −1 x
y
O
x
y
O
y=cos x
x
y
O
y=cos−1
x
1
π
2
π
−1
19
Figure 2 The graphs of the six trigonometric functions and their inverse functions
x
y
O
y=tan x
x
y
O
y=tan−1
x
x
y
O
x
y
O
x
y
O
y=csc x
x
y
O
y=csc−1
x
y=sec x x
y
O
1 −1
y=sec−1
x
2
π
1 −1
2
π
−
y=cot x
x
y
O
y=cot−1
x 2
π
π−
20
3.4.2. Basic Integration by Trigonometric Substitution
Trigonometric substitutions are useful in finding integrals involving 22xa − , 22
xa + and 22ax − ,
where a is a constant. The following examples will illustrate how we use the substitution to find
these integrals.
Type A: Integrals involving 22xa −
We use the substitution θsinax = , where 22
πθ
π<<− .
Example 19
Find ∫−
dx
x2
25
1.
Solution:
Let θsin5=x , then θθddx cos5= .
Cx
C
d
d
d
ddx
x
+=
+=
=
=
=
⋅
−
=
−
−
∫
∫
∫
∫∫
5sin
cos
cos
cos5
cos5
cos5)sin5(25
1
25
1
1
2
22
θ
θ
θθ
θ
θ
θ
θ
θθ
θ
Notes:
(1) In the range 22
πθ
π<<− , θcos is positive. As a result, θθ coscos
2= .
(Normally, we do not have to specify the range of θ in our calculations.)
(2) We can also use the substitution θcos5=x as follows:
Let θcos5=x , then θθddx sin5−= .
Cx
C
d
d
d
ddx
x
+−=
+−=
−=
−=
−=
−⋅
−
=
−
−
∫
∫
∫
∫∫
5cos
sin
sin
sin5
sin5
)sin5()cos5(25
1
25
1
1
2
22
θ
θ
θθ
θ
θ
θ
θ
θθ
θ
21
Example 20
Find ∫ − dxx2
9 .
Solution:
Let θsin3=x , then θθddx cos3= .
C
d
d
d
ddxx
++=
+=
=
−=
⋅−=−
∫
∫
∫
∫∫
θθ
θθ
θθ
θθθ
θθθ
2sin4
9
2
9
)2cos1(2
9
cos9
cossin19
cos3)sin3(99
2
2
22
Now, we need to express the result in terms of x.
3sin
1 x−
=θ
θθθ cossin22sin =
[We construct a right-angled triangle as shown on the right such that3
sinx
=θ .]
[Length of adjacent side = 222
93 xx −=− ]
[Then, we have3
9cos
2x−
=θ .]
9
92
3
9
322sin
22xxxx −
=
−
=θ
Cxxx
Cxxx
dxx
+−+=
+
−+=−
−
−
∫
21
2
12
92
1
3sin
2
9
9
92
4
9
3sin
2
99
Notes:
(1) Sometimes, we may need to further perform integration of trigonometric functions after using
a trigonometric substitution.
(2) The steps of finding θcos (enclosed by [ ]) can be skipped, as long as the right-angled triangle
(with the lengths of three sides clearly indicated) is clearly drawn.
θ
x 3
29 x−
22
Type B: Integrals involving 22xa +
We use the substitution θtanax = , where 22
πθ
π<<− .
Example 21
It is given that Cxxxdx ++=∫ tanseclnsec . (Refer to Example 12 for details)
Find (a) ∫+
dxx 3
1
2 (b) ∫
+
dx
x 3
1
2
.
Solution:
(a) Let θtan3=x , then θθddx2
sec3= .
1
1
1
2
2
2
2
2
2
2
3tan
3
3
3
3
sec
sec
3
3
)1(tan3
sec3
3)tan3(
sec3
3
1
Cx
C
d
d
ddxx
+=
+=
=
+
=
+
=
+
−
∫
∫
∫∫
θ
θθ
θ
θθ
θ
θ
θ
θ
(b) Let θtan3=x , then θθddx2
sec3= .
3ln''' where,''3ln
'3ln3ln
'33
3ln
'tansecln
sec
sec
sec
1tan3
sec3
3)tan3(
sec3
3
1
2
2
2
2
2
2
2
2
2
−=+++=
+−++=
+++
=
++=
=
=
+⋅
=
+
=
+
∫
∫
∫
∫∫
CCCxx
Cxx
Cxx
C
d
d
d
ddx
x
θθ
θθ
θθ
θ
θ
θ
θ
θ
θ
θ
Note:
As 22
πθ
π<<− , θθ secsec
2= , but not θsec− .
θ
x
3
32+x
23
Example 22
Find ∫+
dx
x
x
42
3
by (a) a trigonometric substitution (b) an algebraic substitution.
Solution:
(a) Let θtan2=x , then θθddx2
sec2= .
Cxx
Cxx
Cxx
C
d
d
d
ddx
x
x
+−+=
++−+=
+
+−
+=
+−=
−=
=
⋅=
⋅+
=+
∫
∫
∫
∫∫
)8(43
1
44)4(3
1
2
48
2
4
3
8
sec8sec3
8
)(sec)1(sec8
sectan8
sec2sec2
tan8
sec24)tan2(
)tan2(
4
22
22
3
2
23
2
3
2
3
2
2
3
2
2
3
2
3
θθ
θθ
θθθ
θθ
θ
θ
θθ
θ
θ
(b) Let 42+= xu , then xdxdu 2= .
Cxx
Cxx
Cuu
duuu
du
u
u
duxu
xdx
x
x
+−+=
++−+=
+−=
−=
⋅−
=
⋅=
+
∫
∫
∫∫
−
)8(43
1
)4(4)4(3
1
43
1
)4(2
1
4
2
1
2
1
4
22
2
1
22
3
2
2
1
2
3
2
1
2
1
2
1
3
2
3
Notes:
An integral can be found by different methods. In this example, we used trigonometric substitution
and algebraic substitution, and we still arrived at the same result.
θ
2
42+x
x
24
Type C: Integrals involving 22ax −
We use the substitution θsecax = , where 2
0π
θ << or2
πθπ −<<− .
Example 23
It is given that Cxxxdx ++=∫ tanseclnsec . (Refer to Example 12 for details)
Find (a) ∫−
dx
xx 9
1
2
(b) ∫−
dx
x 9
1
2
.
Solution:
(a) Let θsec3=x , then θθθ ddx tansec3= .
1
1
2
22
3sec
3
1
3
1
tan
tan
3
1
tan3
tan
9)sec3(sec3
tansec3
9
1
Cx
C
d
d
ddx
xx
+=
+=
=
=
−
=
−
−
∫
∫
∫∫
θ
θθ
θ
θ
θ
θ
θ
θθ
θθ
(b) Let θsec3=x , then θθθ ddx tansec3= .
3ln''' where,''9ln
'3
9
3ln
'tansecln
sec
tan
tansec
9)sec3(
tansec3
9
1
2
2
2
22
−=+−+=
+−
+=
++=
=
=
−
=
−
∫
∫
∫∫
CCCxx
Cxx
C
d
d
ddx
x
θθ
θθ
θ
θ
θθ
θ
θ
θθ
Notes:
In the range 2
0π
θ << or2
πθπ −<<− , θθ tantan
2= , but not θtan− .
θ
3
x 9
2−x
25
3.4.3. Variations of Integration by Trigonometric Substitution
Example 24
Find ∫+
dxx
x
2sin1
cos.
Solution:
Let θtansin =x , then θθdxdx2
seccos = .
Cx
C
d
ddxx
x
+=
+=
=
+
=
+
−
∫
∫∫
)(sintan
tan1
sec
sin1
cos
1
2
2
2
θ
θ
θθ
θ
Example 25 (Method of Completing the Square)
Find ∫−+−
dx
xx 2012
1
2
.
Solution:
We first express 20122
−+− xx in the form of 22 )( bxa −− , where a and b are constants.
22
2
22
2
22
)6(4
)6(16
2
1220
2
1212
20)12(2012
−−=
−−=
+−
+−−=
−−−=−+−
x
x
xx
xxxx
Let θsin46 =−x , then θθddx cos4= .
Cx
C
d
d
dx
x
dx
xx
+−
=
+=
=
−
=
−−
=
−+−
−
∫
∫
∫∫
4
6sin
cos
cos
)sin4(4
cos4
)6(4
1
2012
1
1
22
222
θ
θθ
θ
θ
θ
θ
Notes:
The method of completing the square is used in converting quadratic expressions into the form
required for using trigonometric substitutions. We usually use this method when we encounter a
quadratic denominator, while the numerator is a constant.
26
3.5. Integration by Parts
3.5.1. Introduction
There are still some integrals that cannot be found using the above techniques, e.g. ∫ xdxx sin ,
∫ xdxx ln and ∫ xdxex
cos , etc. which are mostly in the form of the products of two functions.
Here, we need a new technique called integration by parts to deal with these integrals.
Let )(xfu = and )(xgv = be two differentiable functions.
Then, we have ∫∫ −= vduuvudv .
The proof is amazingly simple with the help of the product rule of differentiation.
Proof:
'')( vuuvuvdx
d+=
Integrating both sides w.r.t. x, we have
∫∫
∫∫+=
+=
vduudv
dxvudxuvuv ''
i.e. ∫∫ −= vduuvudv .
The following example illustrates how we can use integration by parts to find an integral.
Example 26
Find ∫ xdxln .
Solution:
Consider ∫ xdxln .
Take xu ln= and xv = .
We have dxx
du1
= and dxdv = .
Cxxx
dxxx
dxx
xxx
vduuv
udvxdx
+−=
−=
⋅−=
−=
=
∫
∫
∫
∫∫
ln
ln
1ln
ln
In fact, it can be written in a simpler way:
Cxxx
dxxx
dxx
xxx
xxdxxxdx
+−=
−=
⋅−=
−=
∫
∫
∫∫
ln
ln
1ln
)(lnlnln
27
3.5.2. Common Integrals involving Integration by Parts
There are many integrals that can be found using integration by parts. Some of them are listed here.
Type A: ∫ xdxxn
sin or ∫ xdxxn
cos
Example 27
Find (a) ∫ xdxx sin (b) ∫ xdxx cos (c) ∫ dxx
x2
sin22 .
Solution:
(a) ∫ xdxx sin
Cxxx
xdxxx
xxd
++−=
+−=
−=
∫
∫
sincos
coscos
)(cos
(b) ∫ xdxx cos
Cxxx
xdxxx
xxd
++=
−=
=
∫
∫
cossin
sinsin
)(sin
(c) ∫ dxx
x2
sin22
( )
Cxxxxxx
xdxxxxx
xxdxxx
xdxx
xdxxdxx
dxxx
++−−=
+−=
+−=
−=
−=
−=
∫
∫
∫
∫∫
∫
sincossin2
1
6
sinsin2
1
6
)(sin2
1sin
2
1
6
)(sin2
1
6
cos2
1
2
1
cos12
1
2
3
2
3
22
3
2
3
22
2
Notes:
Integration by parts can be used repeatedly to lower the power of some integrals. In DSE M2
examination, however, you are required to use this technique at most twice only to find an integral.
Here, xu = , xv cos= ,
dxdu = and xdxdv sin−= .
Here, xu = , xv sin= ,
dxdu = and xdxdv cos= .
Use integration by parts
repeatedly here (as in part
(a)) to find ∫ xdxx sin .
28
Type B: ∫ xdxxn
ln
Example 28
Find ∫ xdxxn
ln for (a) 1−≠n (b) 1−=n .
Solution:
(a) ∫ xdxxn
ln
Cxn
xxn
dxxn
xxn
xdxn
xxn
xxdn
nn
nn
nn
n
+
+
−
+
=
+
−
+
=
+
−
+
=
+
=
++
+
++
+
∫
∫
∫
1
2
1
1
11
1
)1(
1ln
1
1
1
1ln
1
1
)(ln1
1ln
1
1
)(ln1
1
(b) ∫−
xdxx ln1
Cx
xxd
dxx
x
+=
=
=
∫
∫
2
)(ln
)(lnln
ln
2
*Think About*:
Can we find the integral by expressing ∫ xdxxn
ln as ∫−+ )ln(ln1
xxxdxxnn ?
Type C: ∫ dxexxn
Example 29
Find (a) ∫ dxxex (b) ∫ dxex
x22 .
Solution:
(a) ∫ dxxex
Cexe
dxexe
exd
xx
xx
x
+−=
−=
=
∫
∫ )(
29
(b) ∫ dxexx22
Cexeex
dxexeex
exdex
dxxeex
xdeex
edx
xxx
xxx
xx
xx
xx
x
++−=
+−=
−=
−=
−=
=
∫
∫
∫
∫
∫
2222
2222
222
222
2222
22
4
1
2
1
2
1
2
1
2
1
2
1
)(2
1
2
1
2
1
)(2
1
2
1
)(2
1
Type D: ∫ bxdxeax
sin or ∫ bxdxeax
cos
Example 30
Find (a) ∫ xdxex
sin (b) ∫ bxdxeax
cos , where 0, ≠ba .
Solution:
(a) ∫ xdxex
sin
∫
∫
∫
∫
∫
∫
−+−=
−+−=
+−=
+−=
+−=
−=
xdxexexe
exdxexe
xdexe
xdxexe
exdxe
xde
xxx
xxx
xx
xx
xx
x
sinsincos
)(sinsincos
)(sincos
coscos
)(coscos
)(cos
Rearranging terms, we have
CCCxexexdxe
Cxexexdxe
xxx
xxx
2
1' where,'sin
2
1cos
2
1sin
sincossin2
=++−=
++−=
∫
∫
Note: You can try to find the integral by first expressing ∫ xdxex
sin as ∫ )(sin x
exd .
You will get the same result at last!
Use integration by parts
repeatedly here to find
∫ dxxex2 .
30
(b) ∫ bxdxeax
cos
∫
∫
∫
∫
∫
∫
−+=
−+=
+=
+=
−=
=
bxdxea
bbxe
a
bbxe
a
bxdea
bbxe
a
bbxe
a
ebxda
bbxe
a
bxdxea
bbxe
a
bxdea
bxea
ebxda
axaxax
axaxax
axax
axax
axax
ax
cossincos1
)(sinsincos1
)(sincos1
sincos1
)(cos1
cos1
)(cos1
2
2
2
22
2
Rearranging terms, we have
Cba
aCCbxe
ba
bbxe
ba
abxdxe
Cbxea
bbxe
abxdxe
a
ba
axaxax
axaxax
22
2
2222
22
22
' where,'sincoscos
sincos1
cos
+
=+
+
+
+
=
++=
+
∫
∫
3.5.3. Other Examples Involving Integration by Parts
Example 31
(a) Let xy 1sin
−
= . By consideringdy
dx, or otherwise, show that
21
1
xdx
dy
−
= .
(b) Hence, or otherwise, find (i) ∫−
xdx1
sin (ii) ∫−
dxx21 )(sin .
Solution:
(a) From xy 1sin
−
= , we have yx sin= .
Then, we have ydy
dxcos= .
By inverse function rule of differentiation, we also have ydy
dx
dx
dy
cos
11
=
=
−
.
Also note that 221sin1cos xyy −=−= .
Thus, we have2
1
1
xdx
dy
−
= .
31
Notes:
(1) For xy 1sin
−
= , 22
ππ
≤≤− y . In this range, ycos is positive, and hence
yy2
sin1cos −= , but not y2
sin1−− .
(2) The result can also be arrived at by finding ∫−
dx
x2
1
1 using the substitution θsin=x .
(Refer to Example 19 for details)
(b) (i) There are two methods for finding this integral.
Method 1:
Cxxx
xd
x
xx
dx
x
xxx
xxdxxxdx
+−+=
−
−
+=
−
−=
−=
−
−
−
−−−
∫
∫
∫∫
21
2
2
1
2
1
111
1sin
)1(1
1
2
1sin
1sin
)(sinsinsin
Method 2:
Let xu1
sin−
= . Then, we have ux sin= , i.e. ududx cos= .
Cxxx
Cxxx
Cuuu
uduuu
uud
uduuxdx
+−+=
++=
++=
−=
=
=
−
−−−
−
∫
∫
∫∫
21
111
1
1sin
)cos(sin)sin(sin)(sin
cossin
sinsin
)(sin
cossin
(ii) Again, there are two methods for this integral.
Method 1:
Cxxxxx
dx
x
xxxxx
xdxxxxx
xxdxx
dx
x
xxxx
xxdxxdxx
+−−+=
−
−−−+=
−−−+=
−+=
−
−=
−=
−−
−−
−−−
−−
−
−
−−−
∫
∫
∫
∫
∫∫
21sin2)(sin
1
121sin2)(sin
)(sin121sin2)(sin
)1(sin2)(sin
1
sin2)(sin
)(sin)(sin)(sin
2121
2
2
2121
122121
2121
2
1
21
212121
By (a), we have
2
1
1
1)(sin
x
xdx
d
−
=−
.
Reminders:
(1) xx =− )sin(sin 1
(2) We find )cos(sin 1x
− by
constructing the triangle as
shown:
21 x−
x
1
x1
sin−
32
Method 2:
We also let xu1
sin−
= . Then, we have ux sin= , i.e. ududx cos= .
Cxxxxx
Cxxxxx
Cuuuuu
uduuuuu
uuduu
uduuuu
uuduu
udu
uduudxx
+−−+=
+−+=
+−+=
−+=
+=
−=
−=
=
=
−−
−−−−−
−
∫
∫
∫
∫
∫
∫∫
21sin2)(sin
)sin(sin2)cos(sinsin2)sin(sin)(sin
sin2cos2sin
cos2cos2sin
)(cos2sin
sin2sin
)(sinsin
)(sin
cos)(sin
2121
111121
2
2
2
2
22
2
221
Example 32
Find (a) ∫ xdxex
cossin
for π<< x0 (b) ∫ dxx)sin(ln for 0>x .
Solution:
(a) Let xu sin= , then xdxx
du cossin2
1= .
Cex
Ceue
dueue
eud
duue
uduexdxe
x
uu
uu
u
u
ux
+−=
+−=
−=
=
=
⋅=
∫
∫
∫
∫∫
sin
sin
)1sin(2
22
22
)(2
2
2cos
(b) Let xu ln= , then dxx
du1
= .
∫
∫∫=
=
udue
uduxdxx
u
sin
sin)sin(ln
Repeating the steps in Example 30, we have Cueueudueuuu
++−=∫ sin2
1cos
2
1sin .
Cxxxx
Cxexedxxxx
+−=
++−=∫
)cos(ln2
1)sin(ln
2
1
)sin(ln2
1)cos(ln
2
1)sin(ln lnln
33
Example 33 (Finding ∫ xdxxqp
sectan for even p and odd q)
It is given that Cxxxdx ++=∫ tanseclnsec . (Refer to Example 12 for details)
Find (a) ∫ xdx3
sec (b) ∫ xdxx sectan2 (c) ∫ xdxx
32sectan (d) ∫ − dxxx 1
22 .
Solution:
(a) ∫ xdx3
sec
∫
∫
∫
∫∫
∫
−++=
++−=
++=
+=
+=
xdxxxxx
xxxxdxx
xxxxd
xdxxdxx
dxxx
3
2
2
sectanseclntansec
tansecln)(tansectansec
tansecln)(sectan
sectansec
)1(tansec
Rearranging terms, we have
2
3
1
3
tansecln2
1tansec
2
1sec
tanseclntansecsec2
Cxxxxxdx
Cxxxxxdx
+++=
+++=
∫
∫
(b) ∫ xdxx sectan2
3
3
3
2
tansecln2
1tansec
2
1
tanseclntansecln2
1tansec
2
1
secsec
sec)1(sec
Cxxxx
Cxxxxxx
xdxxdx
xdxx
++−=
++−++=
−=
−=
∫∫
∫
Alternatively,
∫ xdxx sectan2
∫
∫∫
∫
∫
∫
−+−=
−−=
−=
−=
=
xdxxxxxx
xdxxdxxxx
xdxxx
xxdxx
xxd
2
2
3
tansectanseclntansec
sectansectansec
sectansec
)(tansectansec
)(sectan
34
Rearranging terms, we have
3
2
2
tansecln2
1tansec
2
1sectan
tanseclntansecsectan2
Cxxxxxdxx
cxxxxxdxx
++−=
++−=
∫
∫
(c) ∫ xdxx32
sectan
∫
∫∫
∫∫
∫∫
∫
−+−+−=
+−−=
+−−=
−=
−=
xdxxxxxxxxxx
xdxxxxxdxxxx
xxdxxxxdxx
xxdxxd
xxdx
323
2323
33
3
2
sectan3tansecln2
1tansec
2
1tansectansec
sectantansecsectan3tansec
)(sectantansec)(sectantansec
)(tansec)(tansec
)(tansec)1(sec
Rearranging terms, we have
5
332
4
332
tansecln8
1tansec
8
1tansec
4
1sectan
tansecln2
1tansec
2
1tansecsectan4
Cxxxxxxxdxx
Cxxxxxxxdxx
++−−=
++−−=
∫
∫
(d) Let θsec=x , then θθθ ddx tansec= .
5
2223
5
3
32
2222
1ln8
11
8
11
4
1
tansecln8
1tansec
8
1tansec
4
1
sectan
tansec1secsec1
Cxxxxxx
C
d
ddxxx
+−+−−−−=
++−−=
=
⋅−=−
∫
∫∫
θθθθθθ
θθθ
θθθθθ
Notes:
(1) The technique of integration by parts followed by rearranging terms is very useful in finding
integrals in the form ∫ xdxxqp
sectan , where p is even and q is odd.
(2) Here is a challenging problem: find ∫ +− dxxx 2042 . Try it out!
[Hint: Make good use of part (a) of Example 33 by using a suitable substitution!]
(For answers, please read the appendix of this note on p.46.)
12−x x
1
(by (b))
θ
35
Example 34 (Reduction Formula)
(a) Prove the reduction formulas in Examples 8 and 17, i.e.
(i) ∫∫−
−= dxxnxxdxxnnn 1)(ln)(ln)(ln for 0≥n ;
(ii) ∫∫−−
−+−= xdx
n
nxx
nxdx
nnn 21sin
1cossin
1sin for 2≥n .
(b) Let ∫+
=nmn
x
dxI
)1(, where m and n are positive integers.
Show that nnmnI
mn
mn
xmn
xI
1
)1(1
−+
+
=+
. Hence, or otherwise, find ∫+
32 )1(x
dx.
Solution:
(a) (i) ∫ dxxn)(ln
∫
∫
∫
−
−
−=
⋅−=
−=
dxxnxx
dxx
xxnxx
xxdxx
nn
nn
nn
1
1
)(ln)(ln
1)(ln)(ln
)(ln)(ln
(ii) ∫ xdxn
sin
∫∫
∫
∫
∫
∫
−−−+−=
−−+−=
−+−=
+−=
−=
−−
−−
−−
−−
−
xdxnxdxnxx
dxxxnxx
xdxxnxx
xxdxx
xxd
nnn
nn
nn
nn
n
sin)1(sin)1(cossin
)sin1(sin)1(cossin
cossin)1(cossin
)(sincoscossin
)(cossin
21
221
221
11
1
Rearranging terms, we have
∫∫
∫∫−−
−−
−+−=
−+−=+−
xdxn
nxx
nxdx
xdxnxxxdxn
nnn
nnn
21
21
sin1
cossin1
sin
sin)1(cossinsin)11(
Notes:
(1) Rearranging terms is a very commonly used technique used in proving reduction formula using
integration by parts.
(2) In proving reduction formulas involving trigonometric functions, the trigonometric identities
1cossin22
=+ xx , xx22
sec1tan =+ and xx22
csc1cot =+ are helpful on most occasions.
36
(b) nI
∫
∫
∫
+
+
−
++
+=
++
+=
+−
+=
dxx
xmn
x
x
dxx
mxxn
x
x
xxd
x
x
nm
m
nm
nm
m
nm
nmnm
1
1
)1
)1()1(
)1(
(
)1(
)1(
1
)1(
1
11
)1(
)1(
1
)1(
1
)1(
+
++
−+
+
=
+
−
+
++
+
= ∫∫
nnnm
nmnm
m
nm
mnImnIx
x
dxx
mndxx
xmn
x
x
Rearranging terms, we have
nnmn
nnmn
Imn
mn
xmn
xI
Imnx
xmnI
1
)1(
)1()1(
1
1
−+
+
=
−+
+
=
+
+
First, we take 2=m and 2=n .
∫∫+
−+
+
=
+222232 )1()2)(2(
1)2)(2(
)1)(2)(2()1( x
dx
x
x
x
dx
For ∫+
22 )1(x
dx, take 2=m and 1=n .
∫
∫∫
++
++
+=
+
−+
++
+=
+
18
3
)1(8
3
)1(4
1)1)(2(
1)1)(2(
)1)(1)(2(4
3
)1(4)1(
2222
222232
x
dx
x
x
x
x
x
dx
x
x
x
x
x
dx
For ∫+1
2x
dx, let θtan=x , then θθddx
2sec= .
1
1
1
2
2
2
tan
1tan
sec
1
Cx
C
d
d
x
dx
+=
+=
=
+
=
+
−
∫
∫∫
θ
θ
θ
θθ
Thus, we have Cxx
x
x
x
x
dx++
+
+
+
=
+
−
∫1
22232tan
8
3
)1(8
3
)1(4)1(.
Note: The integral ∫+
32 )1(x
dx can also be found by letting θtan=x at the very beginning.
The reduction formula can even be found by differentiation. Try them out!
37
3.6 Integration of Miscellaneous Functions
Some functions can be integrated with a combination of the above techniques.
3.6.1. Finding Integrals in the Form ∫++
++dx
FExDx
CBxAx
2
2
We follow the following steps:
(In the following, all capital letters represent constants.)
(1) Express ∫++
++dx
FExDx
CBxAx
2
2
as ∫∫++
++ dx
FExDx
IHxGdx
2.
(2) Express ∫++
+dx
FExDx
IHx
2 as ∫ ∫
++
+++
++
dxFExDx
KFExDxd
FExDx
J
2
2
2)( .
(3) Find ∫Gdx , ∫ ++
++
)( 2
2FExDxd
FExDx
J and ∫
++
dxFExDx
K
2 one by one.
The integral is then found.
Example 35
Find ∫ ++
−−dx
xx
xx
52
5432
2
.
Solution:
∫
∫∫
∫∫∫
∫∫∫
++
−++−=
+++
−++
++
−=
++
−
++
+−=
++
+−
++
++=
++
−−
dxx
xxx
dxxx
xxdxx
x
dxxx
dxxx
xdx
dxxx
xdx
xx
xxdx
xx
xx
4)1(
110)52ln(53
4)12(
110)52(
52
153
52
110
52
)22(53
52
2010
52
)52(3
52
543
2
2
2
2
2
22
22
2
2
2
For ∫++
dxx 4)1(
12
, let θtan21=+x , then θθddx2
sec2= .
1
1
1
2
2
2
2
1tan
2
1
2
1
2
1
4)tan2(
sec2
4)1(
1
Cx
C
d
ddxx
++
=
+=
=
+
=
++
−
∫
∫∫
θ
θ
θθ
θ
Thus, we have ∫ ++
−++−=
++
−−−
Cx
xxxdxxx
xx
2
1tan5)52ln(53
52
543 12
2
2
.
This step can be replaced by
long division.
x
x
xx
number real allfor 0
4)1(
52
2
2
>
++=
++
Hence, absolute sign is
not required here for
)52ln( 2 ++ xx .
38
Example 36
(a) Let 111
1
23+−
++
+
≡
+ xx
CBx
x
A
x, where A, B and C are constants. Find A, B and C.
(b) Hence, find ∫+
dxx 1
1
3.
Solution:
(a) From the question, we have111
1
23+−
++
+
≡
+ xx
CBx
x
A
x.
1
)()()(
)1)(1(
)1)(()1(
1
1
3
2
2
2
3
+
++−+++≡
+−+
++++−≡
+
x
CAxACBxBA
xxx
xCBxxxA
x
Hence, we have )()()(1 2CAxACBxBA ++−+++≡ .
Comparing the coefficients of like terms, we have
=+
=−+
=+
1
0
0
CA
ACB
BA
.
Solving the above system of linear equations, we have 3
2 and
3
1 ,
3
1=−== CBA .
(b) ∫∫∫+−
−−
+
=
+
dxxx
xdx
xdx
x 1
2
3
1
1
1
3
1
1
1
23
∫∫
∫∫
∫∫
+
−
++−+−
−+=
−+
+−
++−
−−+=
+−
−−+
+=
dx
x
xxdxx
x
dx
xx
dxxx
xx
dxxx
xxd
x
4
3
2
1
1
2
1)1(
1
1
6
11ln
3
1
2
11
2
1
3
6
1
1
12
6
11ln
3
1
1
42
6
1)1(
1
1
3
1
2
2
2
22
2
2
2
For ∫+
−
dx
x4
3
2
1
1
2, let θtan
2
3
2
1=−x , then θθddx
2sec
2
3= .
1
1
12
2
2
3
12tan
3
2
3
2
3
2
4
3tan
2
3
sec
2
3
4
3
2
1
1C
xCdddx
x
+
−=+==
+
=
+
−
−
∫∫∫ θθθ
θ
θ
Thus, we have Cx
xxxdxx
+
−++−−+=
+−
∫3
12tan
3
1)1ln(
6
11ln
3
1
1
1 12
3.
39
3.6.2. General t-substitution for integrals involving trigonometric functions
There are some integrals that involve trigonometric functions in the denominator and/or numerator
in the integrand, e.g. ∫+
dxxsin1
1 and ∫
++
dxxx 1cossin
1, etc. To find these integrals, we can
use a special method which makes the whole integral be in terms of the variable2
tanx
t = .
Before talking about how we can use this method, we first see how we can express the three basic
trigonometric functions, xsin , xcos and xtan , in terms of 2
tanx
.
Example 37
Let 2
tanx
t = . Express xsin , xcos and xtan in terms of t.
Solution:
2
2
2
2
1
2
2tan1
2tan2
2sec
2tan2
2cos
2cos
2sin
2
2cos
2sin2sin
t
t
x
x
x
x
x
x
x
xxx
+
=
+
=
=
=
=
Notes:
You may also consider constructing a right-angled triangle as shown after finding xtan . In this
case, xsin and xcos can also be found. However, this is NOT considered a good method, as you
can only find xsin and xcos for π<< x0 in this case. Using trigonometric identities would
be a better method as the result is true for all real values of x.
There are also other methods to find out the way to express the three functions
in terms of t. Interested readers can search for more information on the internet.
2
2
2
2
2
2
2
1
1
1
)1(2
1
2tan1
2
1
2sec
2
12
cos2cos
t
t
t
t
x
x
xx
+
−=
+
+−=
−
+
=
−=
−=
2
2
1
2
2tan1
2tan2
tan
t
t
x
x
x
−
=
−
=
40
Example 38
Let ∫+
= dxx
Isin1
1. Using the substitution
2tan
xt = , or otherwise, find I.
Solution:
Let 2
tanx
t = .
dtt
dtx
dx
dxx
dt
2
2
2
1
2
2tan1
2
2sec
2
1
+
=
+
=
=
Cx
Ct
tdt
dttt
dtt
t
t
dxx
I
+
+
−=
+
+
−=
+
+
=
++
=
+
⋅
+
+
=
+
=
∫
∫
∫
∫
12
tan
2
1
2
)1()1(
12
12
2
1
2
1
21
1
sin1
1
2
2
2
2
Notes:
(1) Readers should pay attention to the way we express dx in terms of t.
(2) This integral can be especially found in an easier way:
Cx
xd
x
dxx
x
dxxx
dxxxxx
I
+
+
−=+
+
=
+
=
+
=
++
=
∫
∫∫∫
12
tan
2)1
2(tan
)12
(tan
2
)12
(tan
2sec
)2
cos2
(sin
1
2cos
2cos
2sin2
2sin
1
2
2
2
222
(3) You may also consider:
'sectancos
)(cossec
cos
sin1
sin1
sin1
sin1
12
2
22 ∫ ∫∫∫∫ +−=+=−
=
−
−=
+
Cxxx
xdxdxdx
x
xdx
x
xdx
x
*Think about*:
Can you show that the two answers obtained differ by a constant (1, in this case)?
(by Example 37)
41
Example 39
Find ∫+−
dxxx 5cos3sin2
1.
Solution:
Again, we let 2
tanx
t = , then dtt
dx2
1
2
+
= . (Refer to Example 38)
∫
∫
∫
∫
∫
∫∫
+
+
=
+
+
=
−+
++
=
++=
+++−=
+⋅
+
+
−−
+
=+−
dt
t
dt
t
dt
tt
dttt
dtttt
dtt
t
t
t
tdx
xx
16
3
4
1
1
4
1
4
3
4
14
1
4
11
16
1
2
14
1
124
1
55334
2
1
2
51
13
1
22
1
5cos3sin2
1
2
2
2
2
22
2
2
2
2
Let θtan4
3
4
1=+t , then θθddt
2sec
4
3= .
C
x
Ct
C
d
ddxxx
++
=
++
=
+=
=
+
⋅=
+−
−
−
∫
∫∫
3
12
tan4
tan3
1
3
14tan
3
1
3
1
3
1
16
3tan
4
3
sec
4
3
4
1
5cos3sin2
1
1
1
2
2
θ
θ
θ
θ
θ
Notes:
It is common that we have to use a further substitution to find the integral obtained by letting
2tan
xt = , and the techniques in Section 3.6.1 are useful in most situations.
42
4. Applications of Indefinite Integration
Indefinite integration is useful in both mathematics and daily life applications.
A. Applications in Coordinate Geometry
It is known that )(' xf represents the slope of the tangent of a point (x,y) to the curve )(xfy = .
By integrating )(' xf with respect to x, it is expected that we can obtain the original curve
)(xfy = .
However, is it the case? No!
Integrating )(' xf w.r.t. x, we have
CxfCydxxf +=+=∫ )()('
We observe that there is an extra constant C. In fact, what
we have obtained is a family of curves, which can be
obtained by translating the curve of )(xfy = upwards or
downwards.
We can further observe that all these curves have the same
)(' xf , i.e. the slope of tangent to all curves at point (x,y) is
the same.
In order to find out the equation of a specific curve instead
of a family of curves, we need the coordinates of a specific
point to find out the constant C. This restricts the possibility
of curves to only one curve, but not the family of curves as
mentioned.
Similarly, suppose )('' xf is given, we have
1)(')('' Cxfdxxf +=∫
[ ]211
)()(' CxCxfdxCxf ++=+∫
To find out )(xf from )('' xf , we need to integrate twice.
We also need the coordinates of two points on the curve.
By substituting them into21
)( CxCxfy ++= , we can find out the constants 1
C and 2
C .
Finally, we can find out the original function )(xf .
Refer to the examples on the next page.
x
y
O
Figure 3 The slope of tangent to the
curve at point (x,y) is equal to )(' xf .
Slope = )(' xf
x
y
O
Figure 4 These curves, obtained by
translating )(xfy = upwards or
downwards, have the same )(' xf .
43
Example 40
The slope of any point (x,y) on a curve Q is 2
1 x
x
+
. It is known that Q passes through the origin.
Find the equation of Q.
Solution:
From the question, we have 2
1 x
x
dx
dy
+
= .
Integrating both sides w.r.t. x, we have
Cx
xdx
dxx
xy
++=
+
+
=
+
=
∫
∫
2
2
2
2
1
)1(1
1
2
1
1
Put (0,0) into Cxy ++=2
1 , we have
1
010 2
−=
++=
C
C
Thus, the equation of Q is 11 2−+= xy .
Example 41
For curve Y, it is given that xedx
yd 2
2
2
= . If Y passes through points A(0,1) and B(1,1), find the
equation of curve Y.
Solution:
From xedx
yd 2
2
2
= , by integration, we have 1
22
2
1Cedxe
dx
dyxx
+== ∫ .
By further integration, we have 21
2
1
2
4
1
2
1CxCedxCey xx ++=
+= ∫
Put (0,1) and (1,1) into 21
2
4
1CxCey
x
++= , we get 4
3
2=C and 2
21
4
11 eCC −=+ .
Solving the above system of linear equations, we get 4
12
1
eC
−= and
4
3
2=C .
Thus, the equation of curve Y is 4
3
4
1
4
12
2+
−+= x
eey
x .
44
B. Applications in Physics
Consider an object travelling in a straight line:
We define some terms as follows:
Displacement (s) is the distance of an object from its original position, with the consideration of
direction.
Velocity (v) is the derivative of displacement with respect to time (t).
Acceleration (a) is the derivative of velocity with respect to time.
If we take right hand side as positive direction, If PQ = 3 m, then the displacement of P is 3 m. On
the contrary, if we take left hand side as positive direction, then the displacement of P is −3 m.
Note that 2
2
dt
sd
dt
dva == .
By integrating a or v with respect to t, we can obtain v and s respectively.
Example 42
The velocity of a particle travelling along a straight line is given by ttv 25)( −= , where )(tv is the
velocity (in m/s) and t is time (in second). It is known that the particle passes through its original
position after 4 seconds. Take right hand side as positive direction.
(a) Find the displacement of the particle after t seconds.
(b) What is the displacement of the particle when it is momentarily at rest?
Solution:
(a) Let the displacement of the particle after t seconds be )(ts m.
Then Cttdttdttvts +−=−== ∫∫25)25()()( .
When 4=t , 0)( =ts .
4
0)4()4(52
−=
=+−
C
C
Thus, the displacement of the particle is ( 452
−+− tt ) m.
(b) The particle is momentarily at rest when 0)( =tv , i.e. 2
5=t .
The displacement of particle = m 4
94
2
55
2
52
=−
+
−
P Q
45
C. Other Applications
Other than in physics, in many situations, if the rate of change is given, we can find out the original
quantity by integrating the rate with respect to time.
Example 43
The figure on the right shows an inverted circular cone with base radius 30
cm and height 40 cm. At the beginning, the cone was full of water. A small
hole was then drilled at the apex of the cone and water is flowing out of
the cone at a rate of )12( +t cm3/s, where t is time in second.
(a) Express the volume of water remaining in the cone after t seconds.
(b) After how many seconds will the cone be empty?
Correct your answer to the nearest second.
Solution:
(a) Let the volume of water remaining in the cone be V cm3.
Then, we have
Ctt
dttV
tdt
dV
+−−=
+−=
+−=
∫2
)12(
)12(
When 0=t , we have 32 cm 12000)40()30(3
1ππ ==V .
Thus, the volume of water remaining in the cone after t seconds is )12000( 2tt −−π cm
3.
(b) When the cone is empty, 0=V .
194
)0 rejected,(2
1480001or
2
1480001
)1(2
)12000)(1(4)1(1
012000
012000
2
2
2
≈
>+−−++−
=
−−±−=
=−+
=−−
t
t
tt
tt
Q
ππ
π
π
π
Thus, the required time is 194 seconds.
More applications of integration will be found in the topic ‘Definite integrals’.
The writer would like to thank Mr. Yue Kwok Choy for his kind reading and checking of this
article.
-END-
Figure 5 An inverted circular cone
Negative sign is added
since the volume of
water is decreasing.
46
Appendix 1
Proof of Method of Substitution in Integration
To prove:
Let )( xgu = . Then, we have ∫ ∫= duufdxxgxgf )()('))(( .
Consider the function ))(( xgFy = , where ))(( xgF is a primitive function of ))(( xgf .
Then )('))(()(' ))((' xgxgfxgxgFdx
dy== .
Integrating both sides w.r.t. x, we have dxxgxgfy ∫= )(' ))(( .
But )()(' ufuFdu
dy == .
Integrating both sides w.r.t. u, we have ∫= duufy )( .
Then, we have ∫ ∫= duufdxxgxgf )()('))(( .
This completes this proof.
Appendix 2
Finding ∫ +− dxxx 2042
∫∫ +−=+− dxxdxxx222 4)2(204
Now let θtan42 =−x , then θθddx2
sec4= .
C'xxxxxx
Cxxxxxx
Cxxxx
d
d
dxxdxxx
+−++−++−−=
+−
++−
+
−
+−=
+++=
=
⋅+=
+−=+−
∫
∫
∫∫
2204ln8204)2(2
1
4
2
4
204ln8
4
2
4
2048
33(a)) Example(by tansecln8tansec8
sec16
sec44)tan4(
4)2(204
22
22
3
222
222
θθ
θθθ