Chapter 2.

Methods of Integration

Basic Integrals

2.1 Note: We have the following list of Basic Integrals

xp+1
Z Z
p
x dx = + c , for p 6= −1 sec2 x dx = tan x + c
p+1
Z Z
dx
= ln |x| + c sec x tan x dx = sec x + c
x
Z Z
x x
e dx = e + c tan x dx = ln | sec x| + c
ax
Z Z
x
a dx = +c sec x dx = ln | sec x + tan x| + c
ln a
Z Z
dx
ln x dx = x ln x − x + c 2
= tan−1 x + c
1+x
Z Z
dx
sin x dx = − cos x + c √ = sin−1 x + c
1−x 2
Z Z
dx
cos x dx = sin x + c √ = sec−1 x + c
2
x x −1
Proof: Each of these
Z equalities is easy to verify by taking the derivative of the right side. For
d
example, we have ln x dx = x ln x−x+c since (x ln x−x) = 1·ln x+x· x1 −1 = ln x, and
Z dx
d sec x tan x
we have tan x dx = ln | sec x| + c since (ln | sec x|) = = tan x, and we have
dx sec x
sec x tan x + sec2 x
Z
d
sec x dx = ln | sec x + tan x| + c since (ln | sec x + tan x|) = = sec x.
dx sec x + tan x
Z 4 2
x −5
2.2 Example: Find √ dx.
1 x
Solution: By the Fundamental Theorem of Calculus and Linearity, we have
Z 4 2 Z 4
x −5 h i4
x3/2 − 5x−1/2 dx = 25 x5/2 − 10x1/2 = 64 − 20 − 52 − 10 = 12



√ dx = 5 5 .
1 x 1 1
√
Z π/3
2.3 Example: Find sin 2x + cos 3x dx.
π/6
d
Solution: We find antiderivatives for sin 2x and cos 3x. Since dx cos 2x = −2 sin 2x we have
d 1 d d 1



dx − 2 cos 2x = sin 2x and since dx sin 3x = 3 cos 3x we have dx 3 sin 3x cos 3x, and so
Z π/3 h iπ/3
1 1
= 41 + 0 − − 14 + 13 = 16 .



sin 2x + cos 3x dx = − 2 cos 2x + 3 sin 3x
π/6 π/6

1
Substitution
2.4 Theorem: (Substitution, or Change of Variables) Let u = g(x) be differentiable on
an interval and let f (u) be continuous on the range of g(x). Then
Z Z
0
f (g(x))g (x) dx = f (u) du

and Z b Z g(b)
0
f (g(x))g (x) dx = f (u) du .
x=a u=g(a)
Z
0
Proof: Let F (u) be an antiderivative of f (u) so F (u) = f (u) and f (u) du = F (u) + c.
d
Then from the Chain Rule, we have F (g(x)) = F 0 (g(x))g 0 (x) = f (g(x))g 0 (x), and so
dx
Z Z
0
f (g(x))g (x) dx = F (g(x)) + c = F (u) + c = f (u) du

and Z b h ib
f (g(x))g 0 (x) dx = F (g(x)) = F (g(b)) − F (g(a))
x=a x=a
h ig(b) Z g(b)
= F (u) = f (u) du .
u=g(a) u=g(a)

2.5 Notation: For u = g(x) we write du = g 0 (x) dx. More generally, for f (u) = g(x) we
write f 0 (u) du = g 0 (x) dx. This notation makes the above theorem easy to remember and
to apply.
√
Z
2.6 Example: Find 2x + 3 dx.

Solution: Make the substitution u = 2x + 3 so du = 2dx. Then

√
Z Z
1 1/2
2x + 3 dx = 2u du = 13 u3/2 + c = 31 (2x + 3)3/2 + c .
√
(In applying the Substitution Rule, we used u = g(x) = 2x + 3 and f (u) = u = u1/2 ,
but the notation du = g 0 (x) dx allows us to avoid explicit mention of the function f (u) in
our solution).
Z
2
2.7 Example: Find x ex dx.

Solution: Make the substitution u = x2 so du = 2x dx. Then

Z Z
x2 1 u 1 u 1 x2
x e dx = 2 e du = 2 e + c = 2 e + c.

2
 Z
ln x
2.8 Example: Find dx.
x
1
Solution: Let u = ln x so du = dx. Then
x
Z Z
ln x
dx = u du = 12 u2 + c = 21 (ln x)2 + c .
x
Z
2.9 Example: Find tan x dx.

sin x
Solution: We have tan x = . Let u = cos x so du = − sin x dx. Then
cos x
−du
Z Z Z
sin x dx
tan x dx = = = − ln |u| + c = − ln | cos x| + c = ln | sec x| + c .
cos x u
Z
dx
2.10 Example: Find √ .
x+ x
√
Solution: Let u = x so u2 = x and 2u du = dx. Then
Z Z Z
dx 2u du 2 du
√ = 2
= .
x+ x u +u u+1
Now let v = u + 1 do dv = du. Then
√
Z Z Z
dx 2 du 2
√ = = dv = 2 ln |v| + c = 2 ln |u + 1| + c = 2 ln( x + 1) + c .
x+ x u+1 v
Z 2
x dx
2.11 Example: Find √ .
0 2x2 + 1
Solution: Let u = 2x2 + 1 so du = 4x dx. Then
Z 2 Z 9 1 Z 9
4 du
x dx h i9
1 −1/2 1 1/2
√ = √ = 4 u du = 2 u = 32 − 12 = 1 .
x=0
2
2x + 1 u=1 u 1 1
Z 1
dx
2.12 Example: Find 2
.
0 1 + 3x
√ √
Solution: Let u = 3 x so du = 3 dx. Then
Z 1 Z √3 √1 du i√3
dx 3
h
1 −1
2
= 2
= √3 tan u = √13 π3 = 3√π
3
.
0 1 + 3x 0 1 + u 0

3
Integration by Parts
2.13 Theorem: (Integration by Parts) Let f (x) and g(x) be differentiable in an interval.
Then Z Z
0
f (x)g (x) dx = f (x)g(x) − g(x)f 0 (x) dx

so b
Z b  Z
0 0
f (x)g (x) dx = f (x)g(x) − g(x)f (x) dx .
x=a x=a

d
Proof: By the Product Rule, we have f (x)g(x) = f 0 (x)g(x) + f (x)g 0 (x) and so
dx
Z
f 0 (x)g(x) + f (x)g 0 (x) dx = f (x)g(x) + c ,

which can be rewritten as

Z Z
f (x)g (x) dx = f (x)g(x) − g(x)f 0 (x) dx .
0

(We do not need to include the arbitrary constant c since there is now an integral on both
sides of the equation).
2.14 Notation: If we write u = f (x), du = f 0 (x) dx, v = g(x) and dv = g 0 (x) dx, then
the top formula in the above theorem becomes
Z Z
u dv = uv − v du .

2.15 Note: To find the integral of a polynomial multiplied by an exponential function

or a trigonometric function, try Integrating by parts with u equal to the polynomial (you
may need to integrate by parts repeatedly if the polynomial is of high degree).
To integrate a polynomial (or an algebraic) function times a logarithmic or inverse
trigonometric function, try integrating by parts letting u be the logarithmic or inverse
trigonometric function.
To integrate an exponential function times a sine or cosine function, try integrating
by parts twice, letting u be the exponential function both times.
Z
2.16 Example: Find x sin x dx.

Solution: Integrate by parts using u = x, du = dx, v = − cos x and dv = sin x dx to get

Z Z Z Z
x sin x dx = u dv = uv − v du = −x cos x + cos x dx = −x cos x + sin x + c .

4
 Z
2.17 Example: Find (x2 + 1)e2x dx.

Solution: Integrate by parts using u = x2 + 1, du = 2x dx, v = 12 e2x and dv = e2x dx to

get Z Z Z Z
2 2x
(x + 1)e dx = u dv = uv − v du = 2 (x + 1)e − x e2x dx .
1 2 2x

Z
To find x e2x dx we integrate by parts again, this time using u2 = x, du2 = dx, v2 = 21 e2x
and dv2 = e2x dx to get
Z Z
2 2x
(x + 1)e dx = 2 (x + 1)e − x e2x dx
1 2 2x

 Z 
1 2 2x 1 2x 1 2x
= 2 (x + 1)e − 2 xe − 2e dx
 
1 2 2x 1 2x 1 2x
= 2 (x + 1)e − 2 xe − 4 e +c

= 14 2x2 − 2x + 3 e2x + c


Z
2.18 Example: Find ln x dx.

1
Solution: Integrate by parts using u = ln x, du = dx, v = x and dv = dx to get
x
Z Z
ln x dx = x ln x − 1 dx = x ln x − x + c .
Z 4 √
2.19 Example: Find x ln x dx.
1

1
Solution: Integrate by parts using u = ln x, du = dx, v = 23 x3/2 and dv = x1/2 dx to get
x
Z 4 4  4
√
 Z
2 3/2 2 1/2 2 3/2 4 3/2
x ln x dx = 3 x ln x − 3x dx = 3 x ln x − 9 x
1 1 1
= 16 32 2 4


16 28
3 ln 4 − 9 − 3 ln 1 − 9 = 3 ln 4 − 9 .
Z
2.20 Example: Find ex sin x dx
Z
Solution: Write I = ex sin x dx. Integrate by parts twice, first using u1 = ex , du = ex dx,
v = − cos x and dv = sin x dx, and next using u2 = ex , du2 = ex dx, v2 = sin x and
dv2 = cos x dx to get
Z
I = −e cos x + ex cos x dx
x

 Z 
x x x .
= −e cos x + e sin x − e sin x dx

= −ex cos x + ex sin x − I

Thus 2I = −ex cos x + ex sin x + c and so I = 21 (sin x − cos x)ex + d.

5
 Z
2.21 Example: Let n ≥ 2 be an integer. Find a formula for sinn x dx in terms of
Z Z Z
n−2
sin x dx, and hence find sin x dx and sin4 x dx.
2

Z Z
Solution: Let I = sin x dx = sinn−1 x sin x dx. Integrate by parts using u = sinn−1 x,
n

du = (n − 1) sinn−2 x cos x dx, v = − cos x and dv = sin x dx to get

Z
n−1
I = − sin x cos x + (n − 1) sinn−2 x cos2 x dx
Z
n−1
= − sin x cos x + (n − 1) sinn−2 x(1 − sin2 x) dx
Z
n−1
= − sin x cos x + (n − 1) sinn−2 x dx − (n − 1)I .
Z
n−1
Add (n − 1)I to both sides to get nI = − sin x cos x + (n − 1) sinn−2 x dx, that is
Z Z
n n−1
1
sin x dx = − n sin x cos x + nn−1
sinn−2 x dx .

In particular, when n = 2 we get

Z Z
sin2 x dx = − 12 sin x cos x + 1
2 1 dx = − 21 sin x cos x + 12 x + c

and when n = 4 we get

Z Z
sin4 x dx = − 14 sin3 x cos x + 3
4 sin2 x dx = − 14 sin3 x cos x − 3
8 sin x cos x + 38 x + c .
Z
2.22 Example: Let n ≥ 2 be an integer. Find a formula for secn x dx in terms of
Z Z
n−2
sec x dx, and hence find sec3 x dx.
Z Z
n
Solution: Let I = sec x dx = secn−2 x sec2 x dx. Using Integrate by Parts with
u = secn−2 x, du = (n − 2) secn−2 x tan x dx, v = tan x and dv = sec2 x dx, we obtain
Z
n−2
I = sec x tan x − (n − 2) secn−2 x tan2 x dx
Z
n−2
= sec x tan x − (n − 2) secn−2 x(sec2 x − 1) dx
Z
n−2
= sec x tan x − (n − 2)I + (n − 2) secn−2 x dx
Z
Add (n − 2)I to both sides to get (n − 1)I = secn−2
x tan x + (n − 2) secn−2 x dx, that is
Z Z
n 1
sec x dx = n−1 sec n−2
x tan x + n−1 secn−2 x dx .
n−2

In particular, when n = 3 we get

Z Z
sec3 x dx = 12 sec x tan x + 1 1 1
 
2 sec x dx = 2 sec x tan x + 2 ln  sec x + tan x + c

6
Trigonometric Integrals
Z
2.23 Note: To find f (sin x) cos2n+1 x dx, write cos2n+1 x = (1 − sin2 x)n cos x then try
the substitution
Z u = sin x, du = cos x dx.
To find f (cos x) sin2n+1 x dx, write sin2n+1 x = (1 − cos2 x)n sin x then try the
substitution Zu = cos x, du = − sin x dx.
To find sin2m x cos2n x dx, try using the trigonometric identities sin2 θ = 21 − 12 cos 2θ
and cos2 θ = 12 + 12 cos 2θ. Alternatively, write cos2n x = (1 − sin2 x)n and use the formula
from ExampleZ 2.21.
To find f (tan x) sec2n+2 x dx, write sec2n+2 x = (1 + tan2 x)n sec2 x dx and try the
substitution uZ = tan x, du = sec2 x dx.
(sec2 x − 1)n
To find f (sec x) tan2n+1 x dx, write tan2n+1 x = sec x tan x dx and
sec x
try the substitution
Z u = sec x, du = sec x tan x dx.
To find sec2n+1 x tan2n x dx, write tan2n x = (sec2 x − 1)n and use the formula from
Example 2.22.
π/3
sin3 x
Z
2.24 Example: Find dx.
0 cos2 x
Solution: Make the substitution u = cos x so du = − sin x dx. Then
Z π/3 Z π/3 Z 1/2 Z 1/2
sin3 x (1 − cos2 x) sin x dx (1 − u2 ) du 1
2
dx = 2
= − 2
= − 2 + 1 du
0 cos x 0 cos x 1 u 1 u
h i1/2
= u1 + u = 2 + 12 − (1 + 1) = 12 .


1
Z
2.25 Example: Find sin6 x dx.

Solution: We could use the method of example 2.21, but we choose instead to use the
half-angle formulas. We have
Z π/4 Z π/4 Z π/4
6 1 1

3 1 3 3 2 1 3
sin x dx = 2 − 2 cos 2x dx = 8 − 8 cos 2x + 8 cos 2x − 8 cos 2x dx
0 0 0
Z π/4
1 3 3 1 1 1
1 − sin2 2x cos 2x dx



= 8 − 8 cos 2x + 8 2 + 2 cos 4x − 8
0
Z π/4
= 5
16 − 1
2 cos 2x + 3
16 cos 4x + 1
8 sin2 2x cos 2x dx
0
h iπ/4
= 5
15 x − 1
4 sin 2x + 3
64 sin 4x + 1
48 sin3 2x
0
5π 1 1 5π 11
= 64 − 4 + 48 = 64 − 48 .

7
 Z π/4
2.26 Example: Find tan4 x dx.
0
Solution: Note first that
tan4 x = tan2 x(sec2 x − 1) = tan2 x sec2 x − tan2 x = tan2 x sec2 x − sec2 x + 1 .
Z
To find tan2 x sec2 x dx, make the substitution u = tan θ, du = sec2 θ dθ to get
Z Z
tan x sec x dx = u2 du = 31 u3 + c = 13 tan3 x + c .
2 2

Thus we have
Z π/4 Z π/4
4
tan x dx = tan2 x sec2 x − sec2 x + 1
0 0
h iπ/4
1 3 1 π π 2
= 3 tan x − tan x + x = 3 −1+ 4 = 4 − 3 .
0
π/4
sec4 x
Z
2.27 Example: Find √ dx.
0 tan x + 1
Solution: Make the substitution u = tan x so du = sec2 x dx. Then
Z π/4 Z π/4 Z 1 2
sec4 x (tan2 x + 1) sec2 x dx (u + 1) du
√ dx = √ = √
0 tan x + 1 0 tan x + 1 0 u+1
Now make the substitution v = u + 1 so u = v − 1 and du = dv. Then
Z 1 2 Z 2 Z 2
u +1 (v − 1)2 + 1
√ du = √ dv = v 3/2 − 2 v 1/2 + 2 v −1/2 dv
0 u + 1 1 v 1
h i2 √ √ √

= 25 v 5/2 − 43 v 3/2 + 4 v 1/2 = 2·45 2 − 4·23 2 + 4 2 − 25 − 34 + 4


1
√ √
(24−40+60) 2 6−20+60 44 2−46
= 15) − 15 = 15 .

8
 Z Z Z
2.28 Note: To find sin(ax) sin(bx) dx , cos(ax) cos(bx) dx , or sin(ax) cos(bx) dx,
use one of the identities
cos(A − B) − cos(A + B) = 2 sin A sin B
cos(A − B) + cos(A + B) = 2 cos A cos B
sin(A − B) + sin(A + B) = 2 sin A cos B .
Z π/6
2.29 Example: Find cos 3x cos 2x dx.
0

Solution: Since 2 cos 3x cos 2x = cos(3x − 2x) + cos(3x + 2x) = cos x + cos 5x, we have
Z π/6 Z π/6 h iπ/6
1 1 1
cos 2x cos 3x dx = 2 (cos x+cos 5x) dx = 2 sin x+ 10 sin 5x 1
= 41 + 20 3
= 10 .
0 0 0

2du
2.30 Note: The Weirstrass substitution u = tan x2 , x = 2 tan−1 u, dx =
1 + u2
converts sin x and cos x into rational functions of u: indeed we have sin x2 u
= √1−u2 and
1−u2
cos x2 = √ 1
so that sin x = 2 sin x2 cos x2 =
1−u2
2u
1+u2 and cos x = cos2 x
2 − sin2 x
2 = 1+u2 .
Z
dx
2.31 Example: Find .
1 − cos x
2
Solution: We use the Weirstrass substitution u = tan x2 , dx = 1+u
2 1−u
2 du, and cos x = 1+u2

to get
2
1+u2 du
Z Z Z Z
dx 2 du du 1 x
= 2 = 2 ) − (1 − u2 )
= 2
= − + c = − cot + c.
1 − cos x 1 − 1−u
1+u2
(1 + u u u 2

9
Inverse Trigonometric Substitution
p
2.32 Note: To solve an integral involving a2 + b2 (x + c)2 or 1/(a2 + b2 (x + c)2 ), try the
substitution θ = tan−1 b(x+c)
p
a so that a tan θ = b(x + c), a sec θ = a2 + b2 (x + c)2 and
a sec2 θ dθ = b dx.
For an integral involving a2 − b2 (x + c)2 , try the substitution θ = sin−1 b(x+c)
p
a so
p
that a sin θ = b(x + c), a cos θ = a2 − b2 (x + c)2 and a cos θ dθ = b dx.
For an integral involving b2 (x + c)2 − a2 , try the substitution θ = sec−1 b(x+c)
p
a so
p
2 2 2
that a sec θ = b(x + c), a tan θ = b (x + c) − a and a sec θ tan θ dθ = b dx.
Z 1
dx
2.33 Example: Find 2 3/2
.
0 (4 − 3x )
√ √ √
Solution: Let 2 sin θ = 3 x so 2 cos θ = 4 − 3x2 and 2 cos θ dθ = 3 dx. Then
Z 1 Z π/3 √2 cos θ dθ Z π/3 iπ/3
dx 3 1
h
1
2 )3/2
= 3
= √ sec 2
θ dθ = √ tan θ = 14 .
0 (4 − 3x 0 (2 cos θ) 0
4 3 4 3 0
Z √3
dx
2.34 Example: Find √ .
1 x2 x2 + 3
√ √ √ √
Solution: Let 3 tan θ = x so 3 sec θ = x2 + 3 and 3 sec2 θ dθ = dx, and also let
u = sin θ so du = cos θ dθ. Then
Z √3 Z π/4 √ Z π/4 Z π/4
dx 3 sec2 θ dθ 1 sec θ 1 cos θ dθ
√ = 2
√ = 2 dθ = 2
1 x2 x2 + 3 π/6 3 tan θ 3 sec θ π/6 3 tan θ π/6 3 sin θ
Z 1/√2  1/√2 √ √
1 1 2 2 2− 2
= du = − = − 3 + 3 = 3 .
1/2 3 u2 3u 1/2
Z 4√ 2
x −4
2.35 Example: Find dx.
2 x2
√
Solution: Let 2 sec θ = x so 2 tan θ = x2 − 4 and 2 sec θ tan θ dθ = dx. Then
Z 4√ 2 Z π/3 Z π/3 Z π/3
x −4 tan2 θ sec θ dθ tan2 θ sec2 θ − 1
dx = = dθ = dθ
2 x2 0 sec2 θ 0 sec θ 0 sec θ
Z π/3 h iπ/3 √ √
= sec θ − cos θ dθ = ln | sec θ + tan θ| − sin θ = ln(2 + 3) − 23 .
0 0
Z 3
2.36 Example: Find (4x − x2 )3/2 dx.
2
√
Solution: Let 2 sin θ = x − 2 so 2 cos θ = 4x − x2 and 2 cos θ dθ = dx. Then
Z 3 Z π/6 Z π/6
2 3/2 4
(4x − x ) dx = 16 cos θ dθ = 4 (1 + cos 2θ)2 dθ
2
Z0 0
Z
2
= 4 + 8 cos 2θ + 4 cos 2θ dθ = 4 + 8 cos 2θ + 2 + 2 cos 4θ dθ
h
1
iπ/6 √ √
3
√
9 3
= 6θ + 4 sin 2θ + 2 sin 4θ =π+2 3+ 4 =π+ 4 .
0

10
Partial Fractions
f (x)
2.37 Note: We can find the integral of a rational function as follows:
g(x)
Step 1: use long division to find polynomials q(x) and r(x) with deg r(x) < deg g(x) such
f (x) r(x)
that f (x) = g(x)q(x) + r(x) for all x, and note that = q(x) + so
g(x) g(x)
Z Z
f (x) r(x)
dx = q(x) + dx .
g(x) g(x)
(If deg f (x) < deg g(x) then q(x) = 0 and r(x) = f (x)).
Step 2: factor g(x) into linear and irreducible quadratic factors.
r(x)
Step 3: write as a sum of terms so that for each linear factor (ax + b)k we have the
g(x)
k terms
A1 A2 Ak
+ 2
+ ··· +
(ax + b) (ax + b) (ax + b)k
and for each irreducible quadratic factor (ax2 + bx + c)k we have the k terms
B1 x + C1 B2 x + C 2 Bk x + C k
2
+ 2 2
+ ··· + .
(ax + bx + c) (ax + bx + c) (ax2 + bx + c)k
r(x) r(x)
Writing in this form is called splitting into its partial fractions decomposition.
g(x) g(x)
Step 4: solve the integral.
2.38 Example: If g(x) = x(x − 1)3 (x2 + 2x + 3)2 then in step 3 we would write
r(x) A B C D Ex + F Gx + H
= + + + + + .
g(x) x x − 1 (x − 1)2 (x − 1)3 x2 + 2x + 3 (x2 + 2x + 3)2
and then solve for the various constants.
Z 3
x−7
2.39 Example: Find 2
dx.
2 (x − 1) (x + 2)

x−7 A B C
Solution: In order to get 2
= + 2
+ we need
(x − 1) (x + 2) x − 1 (x − 1) x+2
A(x − 1)(x + 2) + B(x + 2) + C(x − 1)2 = x − 7 .
Equating coefficients gives A + C = 0, A + B − 2C = 1 and −2A + 2B + C = −7. Solving
these three equations gives A = 1, B = −2 and C = −1, and so we have
Z 3 Z 3
x−7 A B C
2
dx = + 2
+
2 (x − 1) (x + 2) 2 x−1 (x − 1) x+2
Z 3 i3
1 2 1 h
2
= − − dx = ln(x − 1) + x−1 − ln(x + 2)
2 x−1 (x − 1)2 x+2 2
8
= (ln 2 + 1 − ln 5) − (2 − ln 4) = ln − 1 .
5

11
 √
3
x4 − x3 + 1
Z
2.40 Example: Find dx.
1 x3 + x
x4 − x3 + 1 −x2 + x + 1
Solution: Use long division of polynomials to show that = x−1+ .
x3 + x x3 + x
2
A Bx + C −x + x + 1
Next, note that to get + 2 = 3
we need A(x2 + 1) + (Bx + C)(x) =
x x +1 x +x
−x2 + x + 1. Equating coefficients gives A + B = −1, C = 1 and A = 1. Solving these
three equations gives A = 1, B = −2 and C = 1. Thus
Z √3 4 3 Z √3
x −x +1 1 2x 1
3
dx = x−1+ − 2 + 2 dx
1 x +x 1 x x +1 x +1
h i √3
1 2 2 −1
= 2 x − x + ln x − ln(x + 1) + tan x
1
√ √ π
= 2 − 3 + ln 3 − ln 4 + 3 − 2 − 1 − ln 2 + π4
3 1



√ √
= 2 − 3 + ln 23 + 12 π
.
Z 2 5
x + x4 − 2x3 − 2x2 − 5x − 25
2.41 Example: Find I = dx.
1 x2 (x2 − 2x + 5)2
Solution: To get
A B Cx + D Ex + F x5 + x4 − 2x3 − 2x2 − 5x − 25
+ 2+ 2 + 2 =
x x x − 2x + 5 (x − 2x + 5)2 x2 (x2 − 2x + 5)2
we need Ax(x2 − 2x + 5)2 + B(x2 − 2x + 5)2 + (Cx + D)(x2 )(x2 − 2x + 5) + (Ex + F )(x2 ) =
x5 + x4 − 2x3 − 2x2 − 5x − 25. Expanding the left hand side then equating coefficients
gives the 5 equations
A + C = 1 , −4A + B − 2C + D = 1 , 14A − 4B + 5C − 2D + E = −2
− 20A + 14B + 5D + F = −2 , 25A − 20B = −5 , 25B = −25
Solving these equations gives A = −1, B = −1, C = 2, D = 2, E = 2 and F = −18, so
Z 2
1 1 2x + 2 2x − 18
I= − − 2+ 2 + 2 dx
1 x x x − 2x + 5 (x − 2x + 5)2
Z 2
1 1 2x − 2 + 4 2x − 2 − 16
= − − 2+ 2 + 2 dx
1 x x x − 2x + 5 (x − 2x + 5)2
Z 2
1 1 2x − 2 4 2x − 2 16
= − − 2+ 2 + 2 + 2 − dx
1 x x x − 2x + 5 x − 2x + 5 (x − 2x + 5)2 (x2 − 2x + 5)2
Z Z
1 1 1
We have dx = ln x + c and 2
dx = − + c. Make the substitution u = x2 − 2x + 5,
x x x
du = (2x − 2) dx to get
(2x − 2) dx
Z Z
du
2
= = ln u + c = ln(x2 − 2x + 5) + c
x − 2x + 5 u
and
(2x − 2) dx −1 −1
Z Z
du
2 2
= 2
= +c= 2 + c.
(x − 2x + 5) u u x − 2x + 5
√
Make the substitution 2 tan θ = x − 1, 2 sec θ = x2 − 2x + 5, 2 sec2 θ dθ = dx to get
4 · 2 sec2 θ dθ
Z Z Z
4 dx −1 x−1


= = 2 dθ = 2θ + c = 2 tan 2 +c
x2 − 2x + 5 (2 sec θ)2

12
and
16 · 2 sec2 θ dθ
Z Z Z Z Z
16 dx 2 dθ 2
= dθ = = 2 cos θ dθ = 1 + cos 2θ dθ
(x − 2x + 5)2
2 (2 sec θ)4 sec2 θ
2(x−1)
1
sin 2θ + c = θ + sin θ cos θ + c = tan−1 x−1


=θ+ 2 2 + x2 −2x+5 + c.
Thus we have
"
1 x−1
I= − ln x + + ln(x2 − 2x + 5) + 2 tan−1
x 2
#2
1 x−1 2(x − 1)
− 2 − tan−1 − 2
x − 2x + 5 2 x − 2x + 5
1
2
x2 − 2x + 5 1

2x − 1 −1 x − 1
= ln + − 2 + tan
x x x − 2x + 5 2 1
5 1 3 −1 1 1



= ln 2 + 2 − 5 + tan 2 − ln 4 + 1 − 4
−1
= ln 58 − 17
20 + tan
1
2 .
sec3 x dx
Z
2.42 Example: Find .
sec x − 1
Solution: Multiply the numerator and denominator by sec x + 1 to get
sec3 x dx sec3 x(sec x + 1) sec4 x + sec3 x sec4 x sec3 x
Z Z Z Z Z
= dx = dx = dx+ dx .
sec x − 1 (sec2 x − 1) tan2 x tan2 x tan2 x
Make the substitution u = tan x, du = sec2 x dx to get
sec4 x (tan2 x + 1) sec2 x dx
Z Z Z 2
u +1
2 dx = 2 = du
tan x tan x u2
Z
1 1
= 1 + 2 du = u − + c = tan x − cot x + c .
u u
Make the substitution v = sin x, dv = cos x dx and integrate by parts to get
sec3 x
Z Z Z Z
dx cos x dx dv
2 dx = 2 = 2 2 =
tan x cos x sin x (1 − sin x) sin x (1 − v 2 ) v 2
Z Z 1 1
1 1 2 2 1
= + dv = + + dv
1 − v2 v2 1 − v 1 + v v2
 
1 1 1 1  1+v  1
= − 2 ln |1 − v| + 2 ln |1 + v| − v + c = 2 ln  1−v  − v + c
 
(1+sin x)2
= 21 ln 1−sin
1+sin x 1  1+sin x 
x − csc x + c = 2 ln (cos x)2 − csc x + c = ln  cos x  − csc x + c .

sec3 x
Z
Thus dx = tan x − cot x + ln | sec x + tan x| − csc x + c.
sec x − 1

13
Approximate Integration
2.43 Definition: Let f be integrable on [a, b]. We can approximate the integral of f on
[a, b] by any Riemann sum
Z b n
∼
X
I= f (x) dx = f (ck )∆k x
a k=1

where a = x0 < x1 < · · · < xn = b, ∆k x = xk − xk−1 and ck ∈ [xk−1 , xk ]. The nth

Left Endpoint Approximation Ln , the nth Right Endpoint Approximation Rn ,
Z b
th
and the n Midpoint Approximation Mn , for the integral I = f (x) dx are the
a
Riemann sums for f obtained by using the partition of [a, b] into n equal sized subintervals
and by choosing ck to be the left endpoint, the right endpoint, or the midpoint of the k th
subinterval [xk−1 , xk ]. We have
n
X b − a 
Ln = f (xk−1 )∆x = f (x0 ) + f (x1 ) + · · · + f (xn−1 )
n
k=1
n
X b − a 
Rn = f (xk )∆x = f (x1 ) + f (x2 ) + · · · + f (xn )
n
k=1
n
X x
k−1+ xk  b − a x0 +x1

x1 +x2

xn−1 +xn


Mn = f ∆x = f 2 +f 2 + ··· + f 2
2 n
k=1
b−a b−a
where xk = a + n k and ∆x = n .

2.44 Definition: Let f be integrable on [a, b]. The Trapezoidal Approximation Tn

Z b
for the integral I = f (x) dx is defined as follows. We use the partition of [a, b] into n
a
equal-sized subintervals, so we let xk = a + b−a b−a
n k and ∆x = n . Let gk be the linear
polynomial with gk (xk−1 ) = f (xk−1 ) and gk (xk ) = f (xk ). Let g be the piecewise-linear
function defined by g(x) = gk (x) for x ∈ [xk−1 , xk ]. We define
Z b
Tn = g(x) dx .
a
Note that Z xk Z xk
f (xk−1 ) + f (xk )
g(x) dx = gk (x) dx = ∆x
xk−1 xk−1 2
(indeed, the integral measures the area of a trapezoid) so we have
n Z xk n
X X f (xk−1 ) + f (xk ) Ln + Rn
Tn = g(x) dx = ∆x =
2 2
k=1 xk−1 k=1
b − a 
= f (x0 ) + 2f (x1 ) + 2f (x2 ) + · · · + 2f (xn−1 ) + f (xn ) .
2n

14
2.45 Definition: Let f be integrable on [a, b]. For an even positive integer n, we define
Z b
the Simpson Approximation Sn for the integral I = f (x) dx as follows. We partition
a
[a, b] into n equal-sized subintervals. Let xk = a+ b−a b−a n
n k and ∆x = n . For k = 1, 2, · · · , 2 ,
let gk be the quadratic polynomial with g(x2k−2 ) = f (x2k−2 ), g(x2k−1 ) = f (x2k−1 ) and
g(x2k ) = f (x2k ). Let g be the piecewise-quadratic function given by g(x) = gk (x) for
x ∈ [x2k−2 , x2k ]. We define
Z b
Sn = g(x) dx .
a

Note that if h(x) = Ax2 + Bx + C is the quadratic polynomial with h(−1) = u, h(0) = v
and h(1) = w, then we must have u = h(−1) = A − B + C, v = h(0) = C and w = h(1) =
A + B + C. Solving these three equations gives A = u−2v+w
2 , B = w−u
2 and Cv so we have
Z 1 Z 1
u−2v+w 2
h(x) dx = 2 x + w−u
2 x + v dx
−1 −1
h i1
u−2v+w w−u
= 6 x3 + 4 x2 + v x
−1
u−2v+w u+4v+w
= 3 + 2v = 3 .
It follows, by shifting and scaling, that
Z x2k
f (x2k−2 ) + 4f (x2k−1 ) + f (x2k )
gk (x) dx = ∆x .
x2k−2 3
Thus
n/2 Z x2k n/2
X X f (x2k−2 ) + 4f (x2k−1 ) + f (x2k )
Sn = g(x) dx = ∆x
x2k−2 3
k=1 k=1
b − a 
= f (x0 ) + 4f (x1 ) + 2f (x2 ) + 4f (x3 ) + · · · + 2f (xn−2 ) + 4f (xn−1 ) + f (xn ) .
3n
2.46 Theorem: (Error Bounds for Approximate Integration) Suppose that the higher
Z b
order derivatives of f exist and are continuous on [a, b]. Let I = f (x) dx. and let
a
Ln , Rn , Tn , Mn and Sn be the left endpoint, right endpoint, midpoint, trapezoidal and
Simpson approximation of I. Then
2
Ln − I  ≤ (b − a) max f 0 (x)
   
2n a≤x≤b
2
Rn − I  ≤ (b − a) max f 0 (x)
   
2n a≤x≤b
  (b − a)3  00 
Tn − I  ≤ max f (x)
12 n2 a≤x≤b
3
Mn − I  ≤ (b − a) max f 00 (x)
   
2
24 n a≤x≤b
5
Sn − I  ≤ (b − a) max f 0000 (x)
   
180 n4 a≤x≤b
Proof: We may assign some proofs as exercises and we may provide some proofs later.

15
 Z 4π/3
2
2.47 Example: Let f (x) = sin x. Find the exact value I = f (x) dx, find the
0
approximations L8 , R8 , M8 , T8 and S8 , and find a bound on the error for each of these
approximations.
Solution: The exact value of the integral is
Z 4π/3 Z 4π/3 h i4π/3 √
2 1 1 1 1 4π 3
I= sin x dx = 2 − 2 cos 2x dx = 2 x − 4 sin 2x = 3 − 8 .
0 0 0
 
When we divide the interval 0, 4π/3 into 8 equal subintervals, the size each of the subin-
tervals is ∆x = π6 and the endpoints of the subintervals are 0, π6 , π3 , π2 , 2π 5π 7π 4π
3 , 6 , π, 6 , 3 .
Thus the approximations are


π π π π 2π 5π 7π






L8 = 6 f (0)+ f 6 + f 3 + f 2 + f 3 + f 6 + f (π)+ f 6
= π6 0 + 14 + 34 + 1 + 43 + 14 + 0 + 14 = 13π


24 ,


R8 = π6 f π6 + f π3 + f π2 + f 2π 5π 7π 4π







3 + f 6 + f (π)+ f 6 + f 3

= π6 41 + 34 + 1 + 34 + 14 + 0 + 14 + 34 = 2π


3 ,
1

29π
T8 = 2 L8 + R8 = 48 ,


M8 = π6 f 12 π
+ f π4 + f 5π 7π 3π 11π 13π 15π








12 + f 12 + f 4 + f 12 + f 12 + f 12
√ √ √ √ √   √ 
= π6 2−4 3 + 12 + 2+4 3 + 2+4 3 + 12 + 2−4 3 + 2−4 3 + 12 = π6 4 − 43 ,


π
f (0)+ 4f π6 + 2f π3 + 4f π2 + 2f 2π 5π 7π 4π







S8 = 18 3 + 4f 6 + 2f (π) + 4f 6 + f 3
π
0 + 1 + 32 + 4 + 32 + 1 + 0 + 1 + 34 = 43π


= 18 72 .
Note that to find the values of f needed for the midpoint approximation M8 , we used the
identity f (x) = sin2 x = 21 − 12 cos 2x. From this same identity, we obtain f 0 (x) = sin 2x
and then f 00 (x) = 2 cos 2x, f 000 (x) = −4 sin 2x and f 0000 (x) = −8 cos 2x. Thus we find that
max f 0 (x) = 1 , max f 00 (x) = 2 and max f 0000 (x) = 8 .
     
0≤x≤4π/3 0≤x≤4π/3 0≤x≤4π/3

The above theorem gives the following error bounds.

2
L8 − I  ≤ (4π/3) · 1 = π2
 
16 9
  (4π/3)2 π2
Rr − I  ≤ ·1=
16 9
3
Tn − I  ≤ (4π/3) · 2 = 8π 3
 
12 · 62 36
3
Mn − I  ≤ (4π/3) · 2 = 4π 3
 
24 · 62 36
5
Sn − I  ≤ (4π/3) · 8 = 27 π 5
 
180 · 64 5·311

16
Improper Integration
2.48 Definition: Suppose that f : [a, b) → R is integrable on every closed interval
contained in [a, b). Then we define the improper integral of f on [a, b) to be
Z b Z t
f = lim− f
a t→b a
provided the limit exists and, when the improper integral exists and is finite, we say
that f is improperly integrable on [a, b), (or that the improper integral of f on [a, b)
converges). In this definition we also allow the case that b = ∞, and then we have
Z ∞ Z t
f = lim f.
a t→∞ a
Similarly, if f : (a, b] → R is integrable on every closed interval in (a, b] then we define the
improper integral of f on (a, b] to be
Z b Z b
f = lim+ f
a t→a t
provided the limit exists, and we say that f is improperly integrable on (a, b] when the
improper integral is finite. In this definition we also allow the case that a = −∞. For a
function f : (a, b) → R, which is integrable on every closed interval in (a, b), we choose a
point c ∈ (a, b), then we define the improper integral of f on (a, b) to be
Z b Z c Z b
f= f+ f
a a c
provided that both of the improper integrals on the right exist and can be added, and we
say that f is improperly integrable on (a, b) when both of the improper integrals on
the right are finite. As an exercise, you should verify that the value of this integral does
not depend on the choice of c.
2.49 Notation: For a function F : (a, b) → R write
h ib−
F (x) + = lim− F (x) − lim+ F (x) .
a x→b x→a

We use similar notation when F : [a, b) → R and when F : (a, b] → R.

2.50 Note: Suppose that f : (a, b) → R is integrable on every closed interval contained
in (a, b) and that F is differentiable with F 0 = f on (a, b). Then
Z b h ib−
f = F (x) + .
a a

A similar result holds for functions defined on half-open intervals [a, b) and (a, b].
Proof: Choose c ∈ (a, b). By the Fundamental Theorem of Calculus we have
Z b Z c Z b Z c Z t
f= f+ f = lim f + lim f
a a c s→a+ s t→b− c



= lim+ F (c) − F (s) + lim− F (t) − F (c)
s→a t→b
h ib−
= lim− F (t) − lim+ F (s) = F (x) + .
t→b s→a a

17
 Z 1 Z 1
dx dx
2.51 Example: Find and find √ .
0 x 0 x
Solution: We have Z 1 i1
dx h
= ln x + = 0 − (−∞) = ∞
0 x 0

and Z 1 h √ i1
dx
√ = 2 x = 2 − 0 = 2.
0 x 0+
Z 1
dx
2.52 Example: Show that converges if and only if p < 1.
0 xp
Solution: The case that p = 1 was dealt with in the previous example. If p > 1 so that
p − 1 > 0 then we have
Z 1  1
dx −1 
1


p
= p−1
= − p−1 − −∞ =∞
0 x (p − 1)x 0+
and if p < 1 so that 1 − p > 0 then we have
Z 1  1−p 1
dx x 
1

1
p
= = 1−p − 0 = 1−p .
0 x 1 − p 0+
Z ∞
dx
2.53 Example: Show that converges if and only if p > 1.
1 xp
Solution: When p = 1 we have
Z ∞ Z ∞ i∞
dx 1 h
= = ln x = ∞ − 0 = ∞.
1 xp 1 x 1

When p > 1 so that p − 1 > 0 we have

Z ∞  ∞
dx −1 
1

1
= = (0) − − p−1 = p−1
1 xp (p − 1)xp−1 1
and if p < 1 so that 1 − p > 0 then we have
Z ∞  1−p ∞
dx x
 1 
= = ∞ − 1−p = ∞ .
1 xp 1−p 1

18
 Z ∞
2.54 Example: Find e−x dx.
0
Solution: We have Z ∞ h i∞
e−x dx = − e−x = 0 − (−1) = 1 .
0 0
Z 1
2.55 Example: Find ln x dx.
0
Solution: We have
Z 1 h i1
ln x dx = x ln x − x = (−1) − (0) = −1 ,
0 0+
1
ln x x
since l’Hôpital’s Rule gives lim+ x ln x = lim+ 1 = lim+ = lim+ −x = 0.
x→0 x→0
x
x→0 − x12 x→0

2.56 Theorem: (Comparison) Let f and g be integrable on closed subintervals of (a, b),
and suppose that 0 ≤ f (x) ≤ g(x) for all x ∈ (a, b). If g is improperly integrable on (a, b)
then so is f and then we have
Z b Z b
f≤ g.
a a
Z b Z b
On the other hand, if f diverges then g diverges, too. A similar result holds for
a a
functions f and g defined on half-open intervals.
Proof: The proof is left as an exercise.
Z π/2 √
2.57 Example: Determine whether sec x dx converges.
0
π √
Solution: For 0 ≤ x < 2 we have cos x ≥ 1 − π2 x so sec x ≤ 1
2
1− π x
hence sec x ≤ √ 1
2
.
1− π x
2
Let u = 1 − π x so that du = − π2 dx. Then
Z π/2 Z 0 i0
1 h
q dx = − π2 u−1/2 = − π u1/2 = π
1
x=0 1 − π2 x u=1

Z π/2
√
which is finite. It follows that sec x dx converges, by comparison.
0
Z ∞
2
2.58 Example: Determine whether e−x dx converges.
0
2 2 1
Solution: For 0 ≤ u we have eu ≥ 1+u, so for 0 ≤ x we have ex ≥ 1+x2 , so e−x ≤ .
1 + x2
Since Z ∞ i∞
dx h
−1 π
= tan x = 2 ,
0 1 + x2 0
Z ∞
2
which is finite, we see that e−x dx converges, by comparison.
0

19
2.59 Theorem: (Estimation) Let f be integrable on closed subintervals of (a, b). If |f | is
improperly integrable on (a, b) then so is f , and then we have
Z 
 b  Z b
f ≤ |f | .
 

 a  a

A similar result holds for functions defined on half-open intervals.

Proof: The proof is left as an exercise.
Z ∞
sin x
2.60 Example: Show that dx converges.
0 x
Z 1 Z ∞
sin x sin x
Solution: We shall show that both of the integrals dx and dx converge.
0 x 1 x
sin x sin x
Since lim = 1, the function f defined by f (0) = 1 and f (x) = for x > 0 is
x→0+ x x
continuous (hence integrable) on [0, 1]. By part 1 of the Fundamental Theorem of Calculus,
Z 1
the function f (x) dx is a continuous function of r for r ∈ [0, 1] and so we have
r
Z 1 Z 1 Z 1 Z 1
sin x sin x
dx = lim+ dx = lim+ f (x) dx = f (x) dx ,
0 x r→0 r x r→0 r 0
Z 1
sin x
which is finite, so dx converges.
0 x
Integrate by parts using u = x1 , du = − x12 dx, v = − sin x and dv = cos x dx to get
Z ∞  ∞ Z ∞ Z ∞
sin x cos x cos x cos x
dx = − − 2
dx = cos(1) − dx .
1 x x 1 1 x 1 x2
 cos x  Z ∞ Z ∞
1 dx  cos x 

Since  2  ≤ 2 and converges, we see that  dx converges too, by
 
x xZ x2 x2

1 1
∞
cos x
comparison. Thus dx also converges by the Estimation Theorem.
1 x2

20