Part 3, Semester 2.

Primitives and Indefinite Integrals *

Leonid Pankratov

April 22, 2023

Science is the Differential Calculus of the mind. Art the Integral Calculus; they may be beautiful
when apart, but are greatest only when combined.
Sir Ronald Ross (13 May 1857 – 16 September 1932).

* File : IND–INT–[Link]

1
Contents
1 Differential of a Function 4

2 Geometric Meaning of Differential 4

3 Properties of Differential 5

4 Invariance of the First Differential Form 6

5 Primitive and Indefinite Integral 6

6 Basic Methods of Finding the Primitive 8

7 Primitives of Rational Functions 11

R
8 Primitives of the Form R(cos x, sin x) dx 14

R
9 Primitives of the Form R(x, y(x)) dx 17

10 Exercises with Solutions 21

2
Table of Basic Integrals
xn+1
Z
xn dx = + C, n ̸= −1; Z
n+1 sec2 x dx = tan x + C;
Z
1
dx = ln |x| + C; Z
x
sec x tan x dx = sec x + C;
Z Z
u dv = uv − vdu Z
a x
dx = arctan + C;
Z a2 + x 2 a
ex dx = ex + C;
Z
a 1 x+a
dx = ln + C;
Z a2 −x 2 2 x−a
1 x
ax dx = a + C;
ln a Z
1 x
√ dx = arcsin + C;
Z a2 −x 2 a
ln x dx = x ln x − x + C;
Z
a x
√ dx = sec−1 + C;
Z
x x2 − a2 a
sin x dx = − cos x + C;
Z
1 x
Z √ dx = cosh−1 + C
x2
−a 2 a
cos x dx = sin x + C; √
= ln(x + x2 − a2 ) + C;
Z
tan x dx = ln | sec x| + C; Z
1 x
√ dx = sinh−1 + C
Z x 2 + a2 a
√
sec x dx = ln | sec x + tan x| + C; = ln(x + x2 + a2 ) + C.

Any of the formulas above is applied on the corresponding interval of the real line R, where the
integrand is well defined. Notice that if we have more than one interval, then the constant C might be
different on different intervals.

3
 The main goal of this part is to introduce the notions of primitive and, with its help, indefinite
integral, and consider some examples showing how to find the indefinite integral. We start this chapter
with some extra facts from the differential calculus. Namely, in the next section we will consider the
notion of the differential of a function. After that we proceed with the definition of the indefinite
integral, and some basic techniques of integration.

1 Differential of a Function
Let us consider a function f = f (x) on [a; b] and assume that it is defined is in a neighborhood of x0 ,
i.e, x0 ∈ (a, b), the independent variable x has increment ∆x = x − x0 . The corresponding increment
of the function is then
f (x) − f (x0 )
.

Definition 1.1 We say that f : [a, b] → R is differentiable at x0 ∈ [a, b] if the increment of the
function can be written in the form

f (x) − f (x0 ) = A · (x − x0 ) + o(|x − x0 |), as x → x0 .

The linear part of the increment A · (x − x0 ) is called the differential of f at x0 .

Note that the differential is uniquely defined. Indeed, the limit
f (x) − f (x0 ) o(|x − x0 |)

lim = lim A + =A
x→x0 x − x0 x→x0 x − x0
is unique, and is in fact the derivative of f at x0 . In this way, the differentiability of a function can be
formulated either in terms of the differential or in terms of the existence of derivative at a given point.

As we have seen, the derivative f ′ (x0 ) characterizes the rate of change of function f at the point
x0 , while the differential df = f ′ (x0 )(x − x0 ) provides the best linear approximation to the increment
of function f (x) − f (x0 ) in a neighborhood of the same point x0 . It is important to note that df
depends on x − x0 , which is the increment of independent variable.
Let us take f (x) = x. Then dx = x − x0 and we can write
df = f ′ (x0 )dx = dx. (1.1)
One can interpret this equality as the derivative of f = f (x) being the ratio of two differentials.

2 Geometric Meaning of Differential

In Figure 1 we illustrate the main notions from the previous section. For brevity, we denote h = x−x0 .
The slope of secant through (x0 , f (x0 )) and (x, f (x)) is (f (x) − f (x0 ))/h, while the slope of tangent
at (x0 , f (x0 )) is equal to the derivative f ′ (x0 ). That is why the differential of f is the best linear
approximation of f (x) which is measured along the straight line y − f (x0 ) = f ′ (x0 )(x − x0 ).

4
 y y = f (x)
f (x)−f (x0 )
y − f (x0 ) = h
(x − x0 )

f (x)
df

y − f (x0 ) = f ′ (x0 )(x − x0 )

f (x0 )

h = x − x0
x0 x x

Figure 1: Geometric meaning of the differential of a function.

3 Properties of Differential
Consider two differentiable functions f = f (x) and g = g(x), and C is a generic constant. The main
properties of the differential are similar to the properties of the derivative and are summarized below.

Properties of Differential

(i) df (x) = f ′ (x)dx;

(ii) d(Cf ) = Cdf ;

(iii) d(f ± g) = df ± dg;

(iv) d(C) = 0;

(v) d(αx + β) = αdx;

(vi) d (f g) = g df + f dg;
 
f g df − f dg
(vii) d = , g ̸= 0
g g2
(viii) y = f (g(x)) =⇒ dy = f ′ (u) u=g(x)
g ′ (x)dx = y ′ (x)dx.

5
Example 3.1 Let us find the differential of functions xn , ln x, and sin x. Taking the derivatives, we
get the following expressions:

dx
d (xn ) = nxn−1 dx, d (ln x) = , d (sin x) = cos xdx.
x

4 Invariance of the First Differential Form

Let us consider a composition of the functions y = f (u) and u = g(x). Then by the chain rule its
derivative has the form
yx′ = yu′ · u′x ,
where the sub-index stands for the variable of differentiation. The differential of y = f (u) is

dy = yu′ du.

The differential of u = g(x) can be written in a similar way

du = u′x dx.

Plugging du in the previous formula, we get

dy = yu′ du = yu′ u′x dx.

Since yx′ = yu′ · u′x , then

dy = yx′ dx.
Thus in the case of a composite function, the differential expression is the same as in the case of
ordinary one. This property is said to be the invariance of the first differential form.

5 Primitive and Indefinite Integral

While modeling physical processes, and more abstract, in the differential calculus, it is important to
find functions solving equations containing their derivatives, the differential equations. The simplest
problem of this kind is the problem of finding a function F (x) that satisfies the equation F ′ (x) =
f (x). Let us study this problem.

Definition 5.1 A function F (x) is said to be a primitive of a function f (x) on an interval if F is

differentiable on this interval and satisfies the equation F ′ (x) = f (x), or equivalently dF (x) =
f (x) dx.

Example 5.2 The function F (x) = x2 + 3 is a primitive of f (x) = 2x since F ′ (x) = 2x = f (x).

6
 1
Example 5.3 The function F (x) = arctan x is a primitive of the function f (x) = on R. In
1 + x2
fact,

′ 1
arctan x = .
1 + x2
1 1
Example 5.4 The function F (x) = arccot is a primitive of f (x) = on the set R \ {0}, since
x 1 + x2
 
1 1
F ′ (x) = −  2 · − 2 = f (x) for x ̸= 0.
1 x
1+
x

It is clear that primitive is not unique. The following proposition holds.

Proposition 5.5 Let F1 (x) and F2 (x) be two primitives of the function f (x) on the same interval


(a, b). Then F1 − F2 (x) is constant on the interval (a, b).

Remark 1 Notice that the hypothesis that F1 (x) and F2 (x) are compared on a connected interval is
important. It is due to the fact that the function which has zero derivative on a union of two disjoint
intervals, does not need to be constant. For example, we can see this from the comparison of Examples
5.3, 5.4, where the derivatives of F1 (x) = arctan x and F2 (x) = arccot x1 are the same on R \ {0}.
However,

1
F1 (x) − F2 (x) = arctan x − arccot = arctan x − arctan x = 0, x > 0,
x
1
F1 (x) − F2 (x) = arctan x − arccot =
x
= arctan x − π − arctan x = −π, x < 0.

Definition 5.6 In the differential calculus, the operation of taking the differential is known as differ-
entiation. The operation of finding a primitive is known as indefinite integration. The mathematical
notation for this operation is the following
Z
f (x) dx. (5.1)
Z
f (x) dx is said to be the indefinite integral of f (x) on a given interval.

7
 We use (5.1) for any primitive of the function f on the interval, that is the indefinite integral is the
collection of all primitive functions.
Z
In (5.1), the sign is known as the indefinite integral sign, f is called the integrand, and
f (x) dx is said to be a differential form.
From Proposition 5.5 we see that if F (x) is any particular primitive of the function f (x) on the
interval, then Z
f (x) dx = F (x) + C. (5.2)

This means that any other primitive, different from F (x), can be obtained from F (x) by adding a
constant.
If F ′ (x) = f (x), then F is a primitive of f on some interval, and by (5.2),
Z
d f (x) dx = dF (x) = F ′ (x)dx = f (x)dx. (5.3)

In addition, Z Z
dF (x) = F ′ (x)dx = F (x) + C. (5.4)

The relations (5.3) and (5.4) show the reciprocity between the differentiation and indefinite inte-
gration. These operations are reciprocally inverse, up to an arbitrary constant C in (5.4).

6 Basic Methods of Finding the Primitive

R
Let F (x) = f (x)dx, and let φ be a continuously differentiable function. The differentiation rules
imply the following relations:

Z Z Z
(i) [αu(x) + βv(x)] dx = α u(x) dx + β v(x) dx + C; (6.1)
Z Z Z
′ ′
(ii) (uv) dx = u (x)v(x) dx + u(x)v ′ (x) dx + C; (6.2)
Z Z
(iii) f (x) dx = f ◦ φ)(t)φ′ (t) dt . (6.3)
t=φ−1 (x)

Formulas (6.1)-(6.3) enable often to reduce the problem of finding a primitive to the calculation
of the primitives for simpler functions. In a (short) table below we give some basic primitives, which
are found by rewriting the table of derivatives of the elementary functions

Remark 2 (Notational convention.) We know that once a primitive is found for a function on some
interval, the other primitives can be found by adding an arbitrary constant. Then, for the sake of
brevity, we will add this arbitrary constant only to the final result.

8
Linearity of the Indefinite Integral. It follows from (6.1) that the primitive of a linear combination
of functions can be found as the linear combination of primitives.

Example 6.1 Integral of a polynomial function:

Z Z Z Z
a0 + a1 x + ... + an x dx = a0 1 dx + a1 x dx + ... + an xn dx
n

1 1
= a0 x + a1 x2 + ... + an xn+1 + C.
2 n+1

Example 6.2
2
√ √
Z  Z   Z Z
1 2 1 2
x+ √ dx = x +2 x+ dx = x dx + 2 x dx+
x x
Z
1 1 4
+ dx = x3 + x3/2 + ln |x| + C.
x 3 3

Example 6.3
Z x Z
2 1
cos dx = (1 + cos x) dx
2 2
Z Z
1 1 1 1
= 1 dx + cos x dx = x + sin x + C.
2 2 2 2

Integration by Parts. Formula (6.2) can be rewritten as

Z Z
u(x)v(x) + C = u(x) dv(x) + v(x) du(x)

or in the following form

Z Z
u(x) dv(x) = u(x)v(x) − v(x) du(x) + C. (6.4)

This means that when we are looking for a primitive of the function u(x)v ′ (x), we can reduce this
problem to finding a primitive for the function v(x)u′ (x). Formula (6.4) is known as the formula of
integration by parts.

Example 6.4 Find the indefinite integral of ln x. We have :

1
 
Z u(x) = ln x du(x) = dx Z
1
ln x dx = 
 x  = x ln x − x · dx

x
dv(x) = dx v(x) = x
Z
= x ln x − 1 dx = x ln x − x + C.

9
Example 6.5 Find the indefinite integral of x2 ex .

In this example we integrate x2 ex by integrating by parts twice. The first integration is as follows
 
2
Z u(x) = x du(x) = 2x dx
J = x2 ex dx = 
 

x x
dv(x) = e dx v(x) = e
Z
= x e − 2 xex dx = x2 ex − 2 J1 ,
2 x
(6.5)

where
Z
J1 = xex dx.

One can find the integral J1 in a similar way. Namely,

 
Z u(x) = x du(x) = dx Z
J1 = xe dx = 
x x x
 = xe − e dx. (6.6)
 
dv(x) = ex dx v(x) = ex

Now it follows from (6.5) and (6.6) that

 Z 
J = x e − 2 J1 = x e − 2 xe − e dx = x2 ex − 2xex + 2ex + C
2 x 2 x x x

= (x2 − 2x + 2) ex + C.

Change of Variable in an Indefinite Integral. Formula (6.3) shows that when we calculate a primi-
tive for the function (f ◦ φ)(t) · φ ′ (t) we can proceed as follows
Z Z Z
′


(f ◦ φ)(t) · φ (t) dt = f φ(t) dφ(t) = f (x) dx = F (x) + C =


= F φ(t) + C.
In other words, we make first the change of variables φ(t) = x in the integrand, and pass to the new
variable x. Then, on the second step, we find the primitive of the function of x. Finally, we return to
the original variable t by using the formula x = φ(t).

Example 6.6

d(t2 + 1)
Z Z Z
t dt 1 1 dx 1 1
2
= 2
= = ln |x| + C = ln(t2 + 1) + C.
1+t 2 1+t 2 x 2 2

10
Example 6.7
d( x2 ) sin x2
Z Z Z  
dx dx x x 2 x
= = = sin cos = cos
sin x 2 sin x2 cos x2 sin x2 cos x2 2 2 cos x2 2

d( x2 )
Z Z Z
h xi du d(tan u)
= x 2 x = u = = 2
=
tan 2 cos 2 2 tan u cos u tan u
h i Z dv
v = tan u = = ln |v| + C = ln | tan u| + C
v
x
= ln tan + C.
2
The properties (i), (ii), and (iii) are the basic properties allowing to combine different functions and
compute more complicated integrals. The remaining part of the chapter is devoted to the integration of
special classes of functions whose primitives can be written as compositions of elementary functions.

7 Primitives of Rational Functions

Consider Z
P (x)
R(x) dx, R(x) = ,
Q(x)
where the integrand R(x) is a ratio of polynomials. We are going to show that such integrals can be
computed in terms of rational functions, the logarithm and arc tangent function.
Since we work on the real axis, it is possible to write such fraction in the following form
 
kj mj
l n
!
P (x) X X a jk
X X b jk x + c jk
= p(x) + 
k
+ . (7.1)
Q(x) j=1 k=1
(x − x j ) j=1 k=1
(x2 + pj x + qj )k

Here p(x) is a polynomial arising from the division of P (x) by Q(x) (evidently in the case when the
degree of P (x) is grater than or equals to the degree of Q(x); ajk , bjk , and cjk are uniquely defined
real numbers; the polynomial Q(x) is given by
Q(x) = (x − x1 )k1 · . . . · (x − xl )kl · (x2 + p1 x + q1 )m1 · . . . · (x2 + pn x + qn )mn .
Formula (7.1) shows that the integration of the fraction R(x) reduces to the integration of several terms
on the right-hand side of this representation. The integration of the polynomial p(x) is discussed in
Example 6.1. Therefore, it remains to calculate the integrals of fractions of the form
Z Z
1 bx + c
J1 = k
dx and J2 = dx, (k ∈ N). (7.2)
(x − a) (x + p x + q)k
2

The integral J1 can be easily found, since

1


 (x − a)−k+1 + C for k ̸= 1;
J1 = −k + 1 (7.3)

ln |x − a| + C for k = 1.


11
 Consider now the integral J2 . First, completing the square, we get
p2 p2
p  
2 2
x + px + q = x + 2 · ·x+ + q− =
2 4 4
p2
 
 p 2
= x+ + q− , (7.4)
2 4
 2

where q − p4 > 0 since the polynomial x2 + px + q has no real roots.
We set
p p2
x + = t and q − = a2
2 4
to obtain
b (x + p2 ) + c − bp
Z Z
bx + c 2
2 k
dx = 2 k
dx
(x + p x + q) (x + p x + q)
b (x + p2 ) + c − bp
Z Z
2 αt + β
= k
dx = dt,
(t2 + a2 )k
h  i
p 2 p2


x+ 2 + q− 4

where
bp
α = b and β = q − .
2
Moreover, we know that
d(t2 + a2 )
Z Z
t 1
dt =
(t2 + a2 )k 2 (t2 + a2 )k
 1 2 2 −k+1
 2(1 − k) (t + a )

 + C for k ̸= 1;
= (7.5)
 1 ln(t2 + a2 ) + C for k = 1.


2
Thus it remains to study the integral
Z
dt
Hk = . (7.6)
(t2 + a2 )k
Integrating by parts we have that
−2kt
 
u(t) = (t2 + a2 )−k du(t) =
Hk = 
 (t2 + a2 )k+1 

dv(t) = dt v(t) = t
Z 2
t t dt
= 2 2 k
+ 2k
(t + a ) (t + a2 )k+1
2
Z 2
t (t + a2 ) − a2
= 2 + 2k dt
(t + a2 )k (t2 + a2 )k+1
t
= 2 + 2kHk − 2ka2 Hk+1 .
(t + a2 )k

12
Thus
t
Hk = + 2kHk − 2ka2 Hk+1 . (7.7)
(t2 + a2 ) k
Now it follows from (7.7) that

1 t (2k − 1)
Hk+1 = · + Hk . (7.8)
2ka2 (t2 + a2 )k 2ka2

Relation (7.8) enables to lower the power k in (7.6). Moreover, it is easy to calculate H1 . Namely, we
have that
d( at )
Z Z  
dt 1 1 t
H1 = 2 2
= t 2 = arctan + C. (7.9)
(t + a ) a (a) + 1 a a
We observe now that the primitive of Hk can be calculated from (7.8) and (7.9).
Thus the following proposition is proved.

P (x)
Proposition 7.1 The primitive of any rational function R(x) = can be represented in terms of
Q(x)
rational functions and the functions ln and arctan.

Example 7.2 Find the indefinite integral

2x2 + 5x + 5
Z
J= dx. (7.10)
(x2 − 1)(x + 2)

First, we observe that the integrand is a proper fraction and the denominator of this fraction can be
represented as (x − 1)(x + 1)(x + 2). Now using the method of unknown coefficients we write

2x2 + 5x + 5 2x2 + 5x + 5 A B C
2
= = + + , (7.11)
(x − 1)(x + 2) (x − 1)(x + 1)(x + 2) x−1 x+1 x+2

where A, B, C are unknown constants.

Consider the right-hand-side of (7.11). It is easy to see that

A B C
+ +
x−1 x+1 x+2
A(x + 1)(x + 2) + B(x − 1)(x + 2) + C(x − 1)(x + 1)
=
(x − 1)(x + 1)(x + 2)
2
(A + B + C)x + (3A + B)x + (2A − 2B − C)
= ,
(x − 1)(x + 1)(x + 2)

and, therefore,

2x2 + 5x + 5 (A + B + C)x2 + (3A + B)x + (2A − 2B − C)

= .
(x2 − 1)(x + 2) (x − 1)(x + 1)(x + 2)

13
We compare the coefficients in front of x2 , x1 , x0 to obtain the following system of linear equations



 A + B + C = 2,



3A + B = 5, (7.12)




 2A − 2B − C = 5.


Let us find the solution to system (7.12). From the second equation in (7.12), we have that

B = 5 − 3A.

Then the third equation in (7.12) along with the second one implies that

2A − 10 + 6A − C = 5 ⇒ C = 8A − 15.

Now, from the first equation in (7.12), we get

6A − 10 = 2 ⇒ A = 2.

Then, it is easy to see that

B = 5 − 3A = 5 − 4 = −1 and C = 8A − 15 = 16 − 15 = 1.

Thus, it follows from (7.10), (7.11) that

2x2 + 5x + 5
Z Z Z Z
1 1 1
J= dx = 2 dx − dx + dx =
(x2 − 1)(x + 2) x−1 x+1 x+2
(x − 1)2 (x + 2)
= 2 ln |x − 1| − ln |x + 1| + ln |x + 2| + C = ln + C.
(x + 1)

R
8 Primitives of the Form R(cos x, sin x) dx
In this section, the rational function is
R(u, v) = P (u, v)/Q(u, v), where P (u, v) and Q(u, v) are
linear combinations of the terms um v n . Here m = 0,Z1, 2, ... and n = 0, 1, ....
There are several methods for computing the integral R(cos x, sin x) dx. We start by considering
the most general one. However, it is important to notice that this approach is not always very efficient.

a. We make the change of variable x

t = tan . (8.13)
2
14
 First, we observe that

cos x cos2 x2 − sin2 x

2
1 − tan2 x
2 1 − t2
cos x = 2 x = = x = ; (8.14)
sin 2 + cos2 x
2
sin2 x2 + cos2 x
2
1 + tan2 2
1 + t2

sin x 2 sin x2 cos x2 2 tan x2 2t

sin x = 2 x = = x = ; (8.15)
sin 2 + cos2 x
2
sin2 x2 + cos2 x2 1 + tan2 2
1 + t2
dx 2 cos2 x2 2 2dt
dt = x ⇒ dx = 2 x x dt = 2 x dt = . (8.16)
2 cos2 2 sin 2 + cos2 2 1 + tan 2 1 + t2
Now, we have that

1 − t2 2t
Z Z  
2dt
R(cos x, sin x) dx = R ,
1 + t2 1 + t2 1 + t2

and our problem is reduced to the integration of a rational function. However, the rational function
obtained here may be very complicated. That is why in some cases we make use of other approaches
in order to find the integral.
Z Z
2 2
b. The integrals of the form R(cos x, sin x) dx can be reduced to r(tan x) dx, where r(t) is a
rational function, by the change of variables

t = tan x.

In fact, in this case

cos2 x cos2 x 1 1
cos2 x = = 2 = 2
= ;
1 2
cos x + sin x 1 + tan x 1 + t2
2 sin2 x sin2 x tan2 x t2
sin x = = = = ;
1 cos2 x + sin2 x 1 + tan2 x 1 + t2
dx dt
dt = 2
, that is, dx = cos2 x dt = .
cos x 1 + t2

c. Consider the integrals of the form

Z Z
R(cos x, sin2 x) dx or R(cos2 x, sin x) dx.

In this case the functions sin x and cos x can be put under the sign of differential and use the change
of variables
t = cos x or t = sin x, respectively.
After such change of variables, the integrals become

R (t, 1 − t2 ) R (1 − t2 , t)
Z Z
− √ dt or √ dt.
1 − t2 1 − t2

15
Example 8.1 Find the indefinite integral
Z
dx
J= dx. (8.17)
3 + sin x
x


In this example we make use of the universal change of variable t = tan 2
. Then from (8.14)-(8.16)
we get
Z Z
1 2 dt dt
J= 2t · =2 .
3 + 1+t2 1 + t2 3t2 + 2t + 3
Consider now the polynomial 3t2 + 2t + 3. We complete the square to have
  
2 2 2 2 1 1 1
3t + 2t + 3 = 3(t + t + 1) = 3 t + 2 · t · + + 1 −
3 3 9 9
" 2 #
1 8
=3 t+ + .
3 9

Then
d t + 13
Z Z

dt 2
J=2 2
=
2
3t + 2t + 3 3 t + 31 + 98
√ !2
 
Z
1 8 2 2 2 2 du
= u = t + ; = =a = . (8.18)
3 9 3 3 u + a2
2

From the table of indefinite integrals we know that

Z
du 1 u
2 2
= arctan + C. (8.19)
a +u a a
√
Then, substituting a = 2 2/3 from (8.18) into (8.19), we have
Z
2 du 2 1 u 2 3 3u
J= 2 2
= · · arctan + C = · √ · arctan √ + C
3 u +a 3 a a 3 2 2 2 2
x


1 3t + 1 1 3 tan 2 + 1
= √ arctan √ + C = √ arctan √ + C.
2 2 2 2 2 2
Example 8.2 Find the indefinite integral
Z
dx
J= . (8.20)
(sin x + cos x)2

We have that
Z Z   Z
dx d(tan x) dt
J= 2 2
= 2
= t = tan x =
cos x · (tan x + 1) (tan x + 1) (t + 1)2
1 1
=− +C =− + C.
(t + 1) (tan x + 1)

16
 R
9 Primitives of the Form R(x, y(x)) dx
Let R(x, y) be a rational function. Consider some particular integrals which are given by
Z
R(x, y(x)) dx,

where y = y(x) is a function of x.

First, it is clear that if we can make a change of variable x = x(t) such that both functions x = x(t)
and y = y(x(t)) are rational ones with respect to t, then x′ (t) is also a rational function and
Z Z
R(x, y(x)) dx = R x(t), y(x(t)) x′ (t) dt.

This means that our problem is reduced to the integration of a rational function.
a. If s
n ax + b
y= (n ∈ N)
αx + β
then setting
ax + b
tn = (9.21)
αx + β
we obtain that
βtn − b
x= and y = t,
a − αtn
and the integrand becomes a rational function.

Example 9.1 Find the integral

Z r
x−1
J= 3
dx. (9.22)
x+1
In this example we make use of the change of variables (9.21). Then we integrate by parts to have

t3 + 1
  3  Z  3 
x−1 t +1 t +1
J= t =
3
, x= , dx = d = td
x+1 1 − t3 1 − t3 1 − t3
 
 u(t) = t du(t) = dt 3 Z 3

=t· t + 1 t +1
=  3  3 − dt
 t +1 t +1  1−t 3 1 − t3
dv(t) = d v(t) =
1 − t3 1 − t3
t3 + 1 t3 − 1 + 2 t3 + 1
Z Z  
2
=t· − dt = t · − −1 + dt
1 − t3 1 − t3 1 − t3 1 − t3
t3 + 1 t3 + 1
Z Z
1 dt
=t· + t − 2 dt = t · + t − 2
1 − t3 1 − t3 1 − t3 (1 − t)(1 + t + t2 )
3
t +1
=t· + t − 2J1 .
1 − t3

17
 In order to find the integral J1 we write the corresponding integrand as a sum of two fractions
1 A Bt + C
2
= + , (9.23)
(1 − t)(1 + t + t ) (1 − t) (1 + t + t2 )
where A, B, C are unknown constants. We rearrange the right-hand side of (9.23), to have
A Bt + C A · (1 + t + t2 ) + (Bt + C) · (1 − t)
+ =
(1 − t) (1 + t + t2 ) (1 − t)(1 + t + t2 )
(A − B)t2 + (A + B − C)t + (A + C)
= .
(1 − t)(1 + t + t2 )
Thus
1 (A − B)t2 + (A + B − C)t + (A + C)
= . (9.24)
(1 − t)(1 + t + t2 ) (1 − t)(1 + t + t2 )
Comparing the coefficients in front of t2 , t1 , t0 in (9.24), we get the following system of linear
equations with respect to A, B, C




 t2 : (A − B) = 0 ⇒ B = A


 1 1 2
t1 : (A + B − C) = 0 ⇒ C = 2A ⇒ A= , B= , C= .

 3 3 3



 t0 : (A + C) = 1 ⇒ 3A = 1


Thus
Z Z
1 t+2
J1 = dt + dt
3(1 − t) 3(1 + t + t2 )
d(1 − t) 1
Z Z
1 t+2
=− + dt
3 (1 − t) 3 (1 + t + t2 )
Z
1 1 t+2 1 1
= − ln |1 − t| + dt = − ln |1 − t| + J2 . (9.25)
3 3 (1 + t + t2 ) 3 3
Let us compute J2 :
(t + 12 ) + 32
Z Z
t+2
J2 = dt = dt
(1 + t + t2 ) (t + 21 )2 + 34
(t + 21 )
Z Z
3 1
= 1 2 3 dt + dt
(t + 2 ) + 4 2 (t + 2 ) + 43
1 2

d (t + 21 )2 + 34
Z
Z
1 3 dt
= 1 2 3 +
2 (t + 2 ) + 4 2 (t + 2 ) + 43
1 2

 2
1 1 3 3 2 2t + 1
= ln t + + + · √ arctan √ + C
2 2 4 2 3 3
2
√

1 1 3 2t + 1
= ln t + + + 3 arctan √ + C. (9.26)
2 2 4 3

18
Now it follows from (9.25) and (9.26) that

1 1
J1 = − ln |1 − t| + J2
3 3
 2
1 1 1 3 1 2t + 1
= − ln |1 − t| + ln t + + + √ arctan √ + C. (9.27)
3 6 2 4 3 3

Taking into account that

t3 + 1
J=t· + t − 2J1 ,
1 − t3
from (9.26), we obtain
2 !
t3 + 1

2 1 1 3
J=t· 3
+ t + ln |1 − t| − ln t+ +
1−t 3 3 2 4
2 2t + 1
− √ arctan √ + C.
3 3

Since
t3 + 1 t3 + 1
 
2t
t· +t=t +1 = ,
1 − t3 1 − t3 1 − t3
then, finally,
 2
2t 2 1 1 3 2 2t + 1
J= 3
+ ln |1 − t| − ln t+ + − √ arctan √ + C.
1−t 3 3 2 4 3 3

b. Consider the case when √

y= ax2 + bx + c,
that is, integrals of the form Z √
R(x, ax2 + bx + c) dx.

Isolating the perfect square in ax2 + bx + c and using a suitable linear change of variable, we reduce
the general case to one of three cases
Z √ Z √ Z √
2
R(t, t + 1) dt, 2
R(t, t − 1) dt, R(t, 1 − t2 ) dt. (9.28)

In order to obtain the integrals of rational functions, one can make use of the following changes of
variables √ √ √
t2 + 1 = tu + 1 or t2 + 1 = tu − 1 or t2 + 1 = t − u;
√ √ √
t2 − 1 = u(t − 1) or t2 − 1 = u(t + 1) or t2 − 1 = t − u;
√ √ √
1 − t2 = u(1 − t) or t2 − 1 = u(1 + t) or t2 − 1 = tu ± 1.

19
These changes of variables were (probably) proposed by Leonhard Euler. Let us show, for example,
that after the first change of variables one can reduce the first integral in (9.28) to the integral of a
rational function. In fact, if
√
t2 + 1 = tu + 1, then t2 + 1 = t2 u2 + 2tu + 1.

From this relation we have that

2u
t=
1 − u2
and then
s s
√ 4u2 4u2 + (1 − u2 )2
t2 + 1 = +1=
(1 − u2 )2 (1 − u2 )2
s
(1 + u2 )2 (1 + u2 )
= = .
(1 − u2 )2 |1 − u2 |
√
Thus t and t2 + 1 can be represented as rational function of u. Therefore, the integral can be reduced
to the integral of a rational function.
The integrals (9.28) can also be reduced to the integrals of rational functions, with the help of
change of variables t = sinh φ, t = cosh φ, and t = sin φ (or t = cos φ), respectively, to
Z Z
R(sinh φ, cosh φ) cosh φ dφ; R(cosh φ, sinh φ) sinh φ dφ;
Z Z
R(sin φ, cos φ) cos φ dφ; − R(cos φ, sin φ) sin φ dφ.

20
10 Exercises with Solutions

Exercise 10.1 Find the indefinite integral

Z
J = x(x + 1)(x − 2) dx.

Solution to Exercise 10.1. Let us represent the integrand in the form

x(x + 1)(x − 2) = x x2 − 2x + x − 2 = x3 − x2 − 2x.

Then
x4 x3
Z Z Z
J= 3
x dx − 2
x dx − 2 x dx = − − x2 + C.
4 3

Exercise 10.2 Find the indefinite integral

Z r q
√
J= x x x dx.

Solution to Exercise 10.2. Let us represent the integrand in the form

r q
√ √ √
q
x x x = x x3/2 = x · x3/4 = x7/8 .

Then
x7/8+1
Z
8
J= x7/8 dx = + C = x15/8 + C.
7/8 + 1 15

Z x
Exercise 10.3 Find the indefinite integral J = sin2 dx.
2
Solution to Exercise 10.3. Taking into account that

cos 2α = cos2 α − sin2 α = cos2 α − 1 + cos2 α = 2 cos2 α − 1

we obtain
1 + cos 2α
cos2 α = .
2
Consider now the integral J. We have that
Z h Z Z
2 x 2 x 1 + cos x
 i  
J= 1 − cos dx = x − cos dx = x − dx
2 2 2
x sin x x sin x
=x− − +C = − + C.
2 2 2 2

21
Exercise 10.4 Find the indefinite integral

6x − 7
Z
J= dx.
3x2 − 7x + 1
Solution to Exercise 10.4. Since the derivative of the denominator is equal to the numerator

′
3x2 − 7x + 1 = 6x − 7,

we obtain
d(3x2 − 7x + 1)
Z  
J= 2
= t = 3x − 7x + 1
3x2 − 7x + 1
Z
dt
= dt = ln |t| + C = ln |3x2 − 7x + 1| + C.
t

Exercise 10.5 Find the indefinite integral

3x − 6
Z
J= √ dx.
x2 − 4x + 5
Solution to Exercise 10.5. Taking into account that

′ 3
x2 − 4x + 5 = 2x − 4 and 3x − 6 = (2x − 4),
2
we get
2x − 4 d(x2 − 4x + 5) h
Z Z
3 3 i
J= √ dx = √ = t = x2 − 4x + 5
2 x2 − 4x + 5 2 x2 − 4x + 5

3
Z
3 t1/2 √ √
= t−1/2 dt = · + C = 3 t + C = x2 − 4x + 5 + C.
2 2 1/2

Exercise 10.6 Find the indefinite integral

Z
J= sin 2x cos 3x dx.

Solution to Exercise 10.6. In this example we make use of the following formula :
1n o
sin α cos β = sin(α − β) + sin(α + β) . (10.1)
2
Then setting
α = 2x and β = 3x,

22
we get
Z Z Z 
1 1
J= (sin 5x − sin x) dx = sin 5x dx − sin x dx
2 2
 Z   
1 1
= sin 5x d(5x) + cos x = u = 5x
2 5
Z
1 1 1 1
= sin u du + cos x = − cos u + cos x + C
10 2 10 2
1 1
= cos x − cos 5x + C.
2 10

Exercise 10.7 Find the indefinite integral

Z
J= arcsin x dx.

Solution to Exercise 10.7. We have

dx 
du(x) = √
"
u(x) = arcsin x 1 − x2 
J= 
dv(x) = dx
v(x) = x

d(1 − x2 )
Z Z
xdx 1
= x arcsin x − √ = x arcsin x + √
1−x 2 2 1 − x2
Z
1 du
= u = 1 − x2 = x arcsin x +
 
√
2 u
√ √
= x arcsin x + u + C = x arcsin x + 1 − x2 + C.

Exercise 10.8 Find the indefinite integral

√
Z
J= x 1 + x dx.

23
Solution to Exercise 10.8. We have
√ √
Z Z Z
J= (x + 1) 1 + x − 1 + x dx = (x + 1) dx − (x + 1)1/2 dx
3/2

t5/2 t3/2
  Z Z
3/2
= t=1+x = t dt − t1/2 dt = − +C
5/2 3/2

2(x + 1)5/2 2(x + 1)3/2

= − + C.
5 3

Exercise 10.9 Find the indefinite integral

Z
dx
J= √ √ dx.
x+ 4x
Solution to Exercise 10.9. We have
" #
x = t4
Z 3 Z 2
t −1+1
Z
dx t dt
J= √ √ dx = = 4 2
= 4 dt
x+ x 4
dx = 4t3 dt t +t t+1

t2
Z Z  
dt
=4 (t − 1) dt + 4 =4 − t + 4 ln |t + 1| + C
t+1 2
√ √ √

= 2 x − 4 4 x + 4 ln 4 x + 1 + C.

Exercise 10.10 Find the indefinite integral

Z Z
dx dx
J1 = and J2 = .
sin x cos x
Solution to Exercise 10.10. Consider, first, the integral J1 . We have that
dx dx dx d(tan x2 )
= x x = = .
sin x 2 sin 2 cos 2 2 tan x2 cos2 x
2
tan x2
Then
d(tan x2 )
Z
x
J1 = x = ln tan + C x ̸= kπ, k ∈ Z.
tan 2 2
Now we turn to the integral J2 . We have
d x + π2
Z Z

dx π x
J2 = = = ln tan + + C,
sin x + π2


cos x 4 2
π
where x ̸= 2
+ kπ, k ∈ Z.

24
Exercise 10.11 Find the indefinite integral
Z
sin x dx
J= .
1 + cos x

Solution to Exercise 10.11. We have

Z   Z
d(cos x) dt
J=− = t = 1 + cos x = − = − ln (1 + cos x) + C,
1 + cos x t

where x ̸= π + 2kπ, k ∈ Z.

Exercise 10.12 Find the indefinite integral

Z
J= x sinh x dx.

Solution to Exercise 10.12. First, we recall that

ex − e−x ex + e−x
sinh x = and cosh x =
2 2
Integrating by parts we obtain
" #
u(x) = x du(x) = dx
Z
J= x sinh x dx =
dv(x) = sinh x dx v(x) = cosh x
Z
= x cosh x − cosh x dx = x cosh x − sinh x + C.

Exercise 10.13 Find the indefinite integral

Z
J= x ln x dx.

Solution to Exercise 10.13. Integrating by parts, we have

du(x) = dx
" #
u(x) = ln x x2
Z Z
x 1
J = x ln x dx = 2
= ln x − x dx
dv(x) = x dx v(x) = x2 2 2

x2 x2
= ln x − + C.
2 4

25
Exercise 10.14 Find the indefinite integral
Z
J= ln2 x dx.

Solution to Exercise 10.14. In this example we integrate by parts twice. At the first step we have
" 2 dx
#
u(x) = ln x du(x) = 2 ln x
Z
x
J = ln2 x dx = (10.2)
dv(x) = dx v(x) = x
Z
= x ln x − 2 ln x dx = x ln2 x − 2J1 .
2
(10.3)

Now, it remains to find the integral

Z
J1 = ln x dx.

To this end, we make use of the integration by parts. Namely,

" #
u(x) = ln x du(x) = dx
Z Z
x
J1 = ln x dx = = x ln x − dx (10.4)
dv(x) = dx v(x) = x

= x ln x − x + C = x(ln x − 1) + C.

Now from (10.2) and (10.4), we have

J = x ln2 x − 2J1 = x ln2 x − 2x(ln x − 1) + C = x ln2 x − 2 ln x + 2 + C.

Exercise 10.15 Find the indefinite integrals :

Z Z
Js = e sin bx dx and Jc = eax cos bx dx.
ax

Solution to Exercise 10.15. Consider, first, the integral Js . Integrating by parts, we get
" #
u(x) = sin bx du(x) = b cos bx dx
Z
Js = eax sin bx dx =
dv(x) = eax dx v(x) = a1 eax
Z
sin bx ax b
= e − eax cos bx dx.
a a
Thus
sin bx ax b
Js = e − Jc . (10.5)
a a
In a similar way one can find the integral Jc . Namely,
cos bx ax b
Jc = e + Js . (10.6)
a a
26
 Therefore, from equations (10.5), (10.6), we obtain the following linear system of equations with
respect to Js , Jc
b sin bx ax

 Js + a Jc = a e ;


(10.7)
 Jc − b Js = cos bx eax .


a a
We multiply the first equation in (10.7) by ab and add it to the second one. We get

b b2 b b cos bx ax
Js + 2 Jc + Jc − Js = 2 sin bx eax + e .
a a a a a
Then
b2
 
b sin bx + a cos bx ax b sin bx + a cos bx ax
1+ 2 Jc = 2
e =⇒ Jc = e .
a a a2 + b 2
Now let us find the integral Js . From the first equation in (10.7) we have:
b sin bx ax b b sin bx + a cos bx ax sin bx ax
Js = − Jc + e =− · e + e
a a a a2 + b 2 a

eax
 2
a2 + b 2
 
b
= 2 · − + sin bx − b cos bx
a + b2 a a

eax

= · a sin bx − b cos x .
a2 + b 2
Thus
a sin bx − b cos x ax
Js = e .
a2 + b 2

27