You are on page 1of 29

Techniques of Integration

In this chapter we will expand our toolkit of integration techniques. At this point
the only technique, other than just recognizing an antiderivative, that we have developed is u-substitution. By the time we are finished it will require some insight
to choose the correct technique for each problem. Sometimes there will be more
than one method that works, though one technique may be much simpler than
another.
Remember that integration reverses differentiation. In particular u-substitution
reverses the chain rule for derivatives. We begin this chapter by exploiting a technique that reverses another of the basic derivative rules, the product rule.

8.1

Integration by Parts: Reversing the Product Rule

Because integration reverses differentiation, every derivative rule can be reversed


to become an integration rule. Sometimes this is easy; other times it is not. In this
section we will reverse the product rule for derivatives. Lets set up the situation.
Suppose that u( x ) and v( x ) are both differentiable functions. Then the product
rule says
d
[u( x )v( x )] = u( x )v0 ( x ) + u0 ( x )v( x ).
dx
A short-hand way to write this is
d
dv
du
(uv) = u + v .
dx
dx
dx

(8.1)

Of course, if we integrate both sides of (8.1), we get


Z

d
(uv) dx =
dx

dv
dx +
dx

du
dx.
dx

On the left side of this equation integration undoes differentiation, and on the
right side we can simplify the notation as we did in substitution problems so that
we end up with
Z
Z
uv =

u dv +

v du.

Solving for the first term on the right side, we find


Z

u dv = uv

v du.

We have used the product rule to develop a new antidifferentiation rule. This
result is sufficiently important that we single it out as a theorem.
THEOREM 8.1 (Integration by Parts Formula). Suppose that u( x ) and v( x ) are both differen-

tiable functions. Then

u dv = uv

v du.

math 131, techniques of integration

trig integrals

Take-Home Message. Heres how to read Theorem 8.1: We can trade one inteR
gral that we cannot do, say u dv for another that we can (might be able) to do,
R
namely, v du. To do this we need to be able to identify a function u and the
derivative of another function v in the original integrand. This technique is especially useful when an integrand contains two functions that are not related to
each other. By the way, it is called integration by parts because u and v are the
parts. Lets take a look at some examples.
EXAMPLE 8.1.1. Consider the integral

x cos( x ) dx.

The function x is unrelated to the function cos( x ). Compare this to


Z

x cos( x2 ) dx.

In this second situation, the function x is related to cos( x2 ). In particular, x is almost the
derivative of the inside function x2 we are off only by a constant. You should recognize
this second problem as a u-substitution problem. But what about the first integral?
SOLUTION. Lets see how to use integration by parts to solve the integral. In the

R
original integral x cos( x ) dx we have to identify a function u and the derivative of
another function v. Lets try
u=x

dv = cos x dx
R
R
The integration by parts formula then says u dv = uv
v du. So we need to
determine v and du from the information we have so far. Getting du is easy; just take
the derivative of u:
u = x ) du = 1dx = dx.
and

Getting v requires us to integrate dv:


v=

dv =

cos x dx = sin x.

Notice that we have not added +c to the solution for v. As well see below, it turns
R
R
out not to matter in this situation. Thus uv = x sin( x ) and v du = sin( x ) dx. Notice
that we are able to do this latter integral! We have exchanged the integral we could
R
R
not determine, x cos( x ) dx, for one that we can, sin( x ) dx.
Substituting all of this into the integration by parts formula produces
Z

u dv = uv

v du )

x cos( x ) dx = x sin( x )

sin( x ) dx

= x sin( x ) + cos( x ) + c.

We can check that this is correct by differentiating our answer:


d
( x sin( x ) + cos( x ) + c) = sin( x ) + x cos( x )
dx

sin( x ) = x cos( x ),

which is the original integrand. Notice how the product rule comes into play in
checking the answer.
What if #1. What if we had chosen the parts differently. Suppose we let u = cos x
and dv = x dx. Then
u = cos x ) du = sin x dx.
and

x2
.
2
Substituting all of this into the integration by parts formula produces
v=

u dv = uv

v du )

dv =

x dx =

x cos( x ) dx =

x2
cos( x )
2

x2
sin( x ) dx
2

math 131, techniques of integration

trig integrals

which we still do not know how to integrate. In fact, if anything, the new integral is
worse than the original one because of the x2 -term. The power of x has increased. In
our original solution, the power of x decreased because we let u be the function x. Its
derivative du = dx appeared in the new integrand (as well as v), so the new integral
was simpler. We will list a few guidelines below that may help in selecting the parts.
What if #2. What if we had added a constant +C to v in the original solution (so that
v + C = sin x + C). Lets see what happens in the general integration by parts formula
when we substitute v + c for v.
Z

u dv = u(v + C )

(v + C ) du = uv + Cu
= uv + Cu
Z

= uv

Z
Z

v du

v du

Cu

C du

v du.

We get the same answer as in the original integration by parts formula because Cu
and Cu end up canceling each other. In most cases it is simpler to just ignore the
+C.
EXAMPLE 8.1.2 (Classic Parts). Determine

xe

dx.

SOLUTION. We have unrelated functions x and e x so integration by parts may be

useful, since our other technique of substitution does not make sense here. We present
the solution in tabular form which shows the parts in the left column and the integration in the right column.
u=x
du = dx

dv = e x dx
R
R
v = dv =
e x dx = e

R
R

u dv = uv
xe

dx =

v du

xe
xe

x
x

e
x

dx

+c

We can check that this is correct by differentiating our answer:


d
dx

xe

+c =

+ xe

+e

= xe

which is the original integrand.


EXAMPLE 8.1.3. Determine

ln x dx.

SOLUTION. Careful, we dont want the derivative of ln x, we want the antiderivative.

This time we have no choice. Since we dont know how to integrate ln x yet, we have
to differentiate it. In other words we want u = ln x and that leaves dv = dx. So
u = ln x
du =

1
x

dx

dv = dx
R
R
v = dv = dx = x

R
R

u dv = uv

v du
R
ln x dx = x ln x
1 dx = x ln x

x+c

We can check that this is correct by differentiating our answer:



d
1
1 = ln x,
( x ln x x + c) = ln x + x
dx
x
which is the original integrand.

Some Tips for Using Integration by Parts


First, remember that this technique is most useful when there are unrelated functions in the integrand. Second, right now you have only two techniques: substitution and parts. Substitution is usually easier to carry out, so first check to see

math 131, techniques of integration

trig integrals

if that technique applies. At this point, if it does not, then parts is likely to be the
method that works. Third, if you choose parts and the integral gets worse after the first substitution, step back and ask yourself whether the parts should be
chosen differently. Be prepared to change; you may not always choose the parts
correctly the first time.
Here are a couple of general principles that will help you select u and dv when
using integration by parts. There really is no substitute for practice, but these may
help you be more efficient.
Try selecting dv as the most complicated portion of the integrand that you can
integrate. . . and let u be the rest of the integrand.
Try letting u be the portion of the integrand whose derivative is simpler than
u, e.g., du might end up being a lower power of x. This is what we did in Example 8.1.1 when we selected u = x. There du = 1 dx was simpler. If we had
chosen u = cos x, then du = sin x dx would just as complicated as u. The new
integrand would not have been simplified.

Typical Scenarios for Integration by Parts


The next few examples illustrate several basic situations where integration by parts
is useful.
EXAMPLE 8.1.4 (Unrelated Parts). Determine

Z p

x ln x dx.

SOLUTION. The two functions in the integrand,

x and ln x are not related to each


other. Parts should immediately come to mind. Since ln x is not simple to integrate
(though we know how to do it from Example 8.1.3), it makes sense to use it as u. So
u = ln x
du =

1
x

dx

p
dv = x dx
R
Rp
v = dv =
x dx = 23 x3/2

R
u dv = uv
v du
Rp
2 3/2
x ln x dx = 3 x
ln x

= 23 x3/2 ln x
= 23 x3/2 ln x

You should check that this is correct by differentiating the answer.

2 3/2 1
x dx
3x
2 1/2
dx
3x
4 3/2
+c
9x

math 131, techniques of integration

8.2

integration by parts, part ii

Doing It Twice: One Good Turn Deserves Another.

There are a couple of situations where using integration by parts twice is just the
ticket. Part way through you should be able to see why this works.
EXAMPLE 8.2.1 (Parts: Reduction). Determine

x2 sin x dx.

SOLUTION. The two functions in the integrand, x2 and sin x are unrelated to each

other. Parts should come to mind. Since sin x is the more complicated piece and can
be integrated, use it as dv. Likewise, x2 becomes simpler when integrated (sin x does
not), so use it as u. So (being careful with signs)
u = x2

dv = sin x dx
R
R
v = dv = sin x dx =

du = 2x dx

cos x

R
R

u dv = uv
x2 sin x dx =

v du
x2 cos x +

2x cos x dx

R
Notice that the new integral 2x cos x dx is not immediately doable but is simpler
than the original one (the power of x is lower) and can be done with parts.
u = 2x
du = 2 dx

dv = cos x dx
R
R
v = dv = cos x dx = sin x

x2 sin x dx =

x2 cos x +

2x cos x dx
R
2 sin x dx

x2 cos x + 2x sin x

x2 cos x + 2x sin x + 2 cos x + c

Check that this is correct by differentiating the answer.

d 2
x cos x + 2x sin x + 2 cos x + c = 2x cos x + x2 sin x + 2 sin x + 2x cos x
dx
= x2 sin x,

2 sin x

which is the original integrand.


EXAMPLE 8.2.2 (Parts: Circular Reasoning). Determine

e x sin x dx.

SOLUTION. The two functions in the integrand, e x and sin x are unrelated to each

other. Think parts! Here it really does not matter which we use as u or dv, though I
would give a slight preference to using u = sin x since it is slightly easier to differentiate sin x than it is to integrate it (where we get a negative sign). So just to be contrary
I will do it the other way. We will use parts twice, as it turns out.
u = ex
du = e x dx

dv = sin x dx
v = cos x

R
e x sin x dx = e x cos x + e x cos x dx
Watch the signs! This is what I meant above.

R
Notice that the new integral e x cos x dx is not a whole lot different than the first one.
Parts is indicated. Important: Continue to set up the parts in the same way: u is still
the exponential function and dv is still the trig function.
u = ex

dv = cos x dx

du = e x dx

v = sin x

R
R

e x sin x dx =

e x cos x + e x sin x

e x sin x dx. Signs!

e x sin x dx appears on both sides of the equation.


R
Solve for it: 2 e x sin x dx = e x (sin x cos x ) + c
R x
Thus, e sin x dx = 12 e x (sin x cos x ) + c

You should check that this is correct by differentiating the answer. Notice what happened here. We never actually did one of the integrals. We kept applying the parts
formula until the problem more or less circled around on itself. In these sorts of problems you must be very careful with the signs attached to the various integrals. These
are great problems to demonstrate your mastering of this technique.

math 131, techniques of integration

integration by parts, part ii

EXAMPLE 8.2.3 (Udoable Becomes Doable). Determine

arcsin x dx. Sometimes we simply

dont know an antiderivative for a familiar function. Parts can be a way to solve the problem (see Example 8.1.3 for another example of this).
SOLUTION. We dont have much choice. If parts applies we must let u = arcsin x

since we do not know how to integrate it. So


u = arcsin x
du =

p 1
1 x2

dv = dx
R
R
v = dv = dx = x

dx

arcsin x dx = x arcsin x

p x
1 x2

dx

Substitution u = 1 x2 so du = 2x dx ) 12 du = x dx
R
R
arcsin x dx = x arcsin x + 12 u 1/2 du Signs!!
p
R
p
arcsin x dx = x arcsin x + u + c = x arcsin x + 1 x2 + c

You should check that this is correct by differentiating the answer. Another great
problems to demonstrate your understanding of integration techniques.
EXAMPLE 8.2.4 (Tricky). Determine

xe2x
dx.
(2x + 1)2

SOLUTION. It does not look like a substitution problem, so it is probably a parts

problem. Here the choice of parts is tricky. But if we keep in mind the suggestion that
we use the most complicated portion of the integrand that you can integrate for dv",
1
then lets use dv = (2x+
dx. Then by the power rule and a mental adjustment,
1)2
v=

(2x + 1)

dx =

1
(2x + 1)
2

1
.
2(2x + 1)

Now u is the rest of the integrand so u = xe2x and


du = (2xe2x + e2x dx = (2x + 1)e2x dx.
Note the factor of 2x + 1. Putting this all together (and keeping track of signs),
Z

xe2x
dx = uv
(2x + 1)2

v du =

xe2x
+
2(2x + 1)

xe2x
+
2(2x + 1)

xe2x
e2x
+
+c
2(2x + 1)
4

(2x + 1)e2x
dx
2(2x + 1)
e2x
dx
2

where we used a mental adjustment at the last step. You should check that this is
correct by differentiating the answer.

8.3 Problems
1. Integral Mix Up: Before working these out, go through and classify each by the technique that you think will apply: substitution, parts, parts twice, or ordinary methods.
Which cant you do yet? The answers are below.
(a)
(d)
(g)
(j)
(m)

Z
Z

2e

px

dx

(b)

x cos x dx

(e)

( x 2 + 1) e x

+3x

sec2 (2x ) dx
Z

1
1

9x2

dx

(h)

cos xesin x dx

(c)

cos(2px ) dx

(f )

( x2 + 1)e x dx

(i)

x
dx
25 + x2
Z
cos x
p
(n)
dx
1 sin2 x
(k)

dx

e x cos x dx

x2 ln x dx

ln x
dx
x

1
dx
1 + 25x2
Z
sin 1 x
p
(o)
dx
1 x2
(l)

math 131, techniques of integration

integration by parts, part ii

1. Some Answers to the Mix-Up Problem: (All +c".)


2e
p

(a)

px

(b) esin x

(d) x sin x + cos x

(e)

1
2p

sin(2px )

(g)

1 x3 +3x
3e

(h) ( x2

(j)

1
2

tan(2x )

(k)

1
3

(m)

arcsin 3x

(c)

2x + 3)e x

(f )

1 x
2 e (cos x + sin x )

(ln x )2
2

(i)

1 3
3x

ln(25 + x2 )

(l)

(n) arcsin(sin x )

(o)

1
5 arctan 5x
1
2
2 (arcsin x )

1
2

1
3

ln x

Parts: Further Examples


We end with a few more examples that involve trig functions and that provide a
segue to our foray into trig integrals more generally.
EXAMPLE 8.3.1 (Oddball). Determine

tions, it is a composition.

cos(ln x ) dx. Careful! This is not a product of func-

SOLUTION. Again we dont have much choice. Its not substitution since the argument1 is u = ln x and du is nowhere in sight. So lets try parts. Since the integrand

consists of a single function that we dont know how to antidifferentiate we must let
u = cos(ln x ). So
u = cos(ln x )
1
x

du =

dv = dx

sin(ln x ) dx

v=x

R
cos(ln x ) dx = x cos(ln x ) + x 1x sin(ln x ) dx
R
= x cos(ln x ) + sin(ln x ) dx

The term argument means the input


to the function

Sign!

R
The new integral sin(ln x ) dx looks much like the old one, so we try parts again and
hope that we can circle around to where we started as in Example 8.2.2.
u = sin(ln x )
du =

1
x

dv = dx

cos(ln x ) dx

v=x

sin(ln x ) dx = x sin(ln x )

= x sin(ln x )

Putting the two parts integrals together:


Z

cos(ln x ) dx = x cos(ln x ) + x sin(ln x )

which gives
2
or

1
x

cos(ln x ) dx

cos(ln x ) dx

Cycles back

cos(ln x ) dx

cos(ln x ) dx = x cos(ln x ) + x sin(ln x )

1
( x cos(ln x ) + x sin(ln x )) + c
2
You should check that this is correct by differentiating the answer.
cos(ln x ) dx =

EXAMPLE 8.3.2 (Trig). Determine


Z

from

sin( x ) cos(4x ) dx. Careful! How does this problem differ

sin(4x ) cos(4x ) dx? The latter we can do by substitution.

SOLUTION. Its not substitution, so we try parts. It is a little easier to integrate sin x

rather than cos(4x ) because of the constants involved. So


u = cos(4x )
du =

4 sin(4x ) dx

dv = sin x, dx
v=

cos x

sin( x ) cos(4x ) dx =

cos( x ) cos(4x )

Parts again! Should circle around.

cos( x )4 sin(4x ) dx

R
The new integral
cos( x )4 sin(4x ) dx looks much like the old one, so we try parts
again and hope that we can circle around to where we started as in Example 8.2.2.

math 131, techniques of integration

integration by parts, part ii

Remember to choose the parts in the same way at each stage.


u = 4 sin(4x )

dv = cos x dx

du = 16 cos(4x ) dx

v = sin x

So
Z

sin( x ) cos(4x ) dx =

cos( x ) cos(4x )

cos( x ) cos(4x )

cos( x )4 sin(4x ) dx
4 sin( x ) sin(4x )

16

sin( x ) cos(4x ) dx .

This gives
15
or

sin( x ) cos(4x ) dx =

cos( x ) cos(4x )

4 sin( x ) sin(4x )

4 sin( x ) sin(4x )
+c
15
Not too bad, but be careful of your signs throughout similar problems. You should
check that this is correct by differentiating the answer.
sin( x ) cos(4x ) dx =

cos( x ) cos(4x )

Integrating Low Powers of the Secant Function.


R
sec x dx using substitution. We saw
Z

Remember that we were able to solve

sec x dx = ln | sec x + tan x | + c.

Of course

sec2 x dx = tan x + c.

EXAMPLE 8.3.3 (Trig). What about

sec3 x dx?

SOLUTION. Following the suggestions for integration by parts, the most complicated

factor in the integrand that we can integrate is dv = sec2 x dx. Working this out we get
u = sec x
du = sec x tan x dx

dv = sec2 x dx
v = tan x

So
Z

sec3 x dx =

sec x sec2 x dx = sec x tan x

= sec x tan x
= sec x tan x

Z
Z
Z

sec x tan2 x dx
sec x (sec2 x
sec3 x

1) dx

sec x dx

Solving for the original integral we get


2
So

sec3 x dx = sec x tan x +


Z

sec3 x dx =

sec x dx = sec x tan x + ln | sec x + tan x |.

sec x tan x + ln | sec x + tan x |


+ c.
2

We will look at some additional trig integrals in the next section.

math 131, techniques of integration

8.4

trig integrals

Trig Integrals

This section is devoted to integrating powers of trig functions. First we examine


powers of sine and cosine functions.

Powers of a Single Trig Function


We begin with four key trig identities that you should memorize that will make
your life and these integrals much simpler.
Four Key Identities.
cos2 u + sin2 u = 1

(so sin2 u = 1

1 + tan2 u = sec2 u

(so tan2 u = sec2 u

sin u =

1
2

cos2 u =

1
2

1
2

cos 2u

1
2

cos 2u

cos2 u or cos2 u = 1

sin2 u)

1)

(Half angle formula).

The half angle formulas are used to integrate sin2 u or cos2 u in the obvious way.
EXAMPLE 8.4.1. Determine

cos2 (8x ) dx.

SOLUTION. Use equation (3) above with u = 8x. Note the use of a mental adjust-

ment.
Z

cos2 (8x ) dx =

1
2

+ 12 cos(16x ) dx = 12 x +

1
32

sin(16x ) + c.

We have already seen how to integrate low powers of the secant and tangent
functions.
R

tan u du = ln | sec u| + c

sec u du = ln | sec u + tan u| + c

R
R

tan2 u du =

sec2 u

1 du = tan u

(Use a key trig identity.)

u+c

sec2 u du = tan u + c

Higher Powers of Trig Functions


When we want to integrate higher powers (powers greater than n
trig function we can make use of integration by parts.
EXAMPLE 8.4.2. Suppose that n

2 is an integer. Determine

2) of a single

sinn x dx.

SOLUTION. Use integration by parts. The key is to write sinn x as sinn 1 x sin x.

sinn x dx =

sinn

x cos x + (n

1)

sinn

x cos x + (n

1)

sinn

x cos x + (n

1)

Z
Z

sinn

x cos2 x dx

sinn

x (1

sinn

sin2 x ) dx
sinn x dx

Combining all the sinn x terms,


n

sinn x dx =

sinn

x cos x + (n

1)

sinn

x dx

u = sinn
du = (n

x
1) sinn

x cos x dx

dv = sin x dx
v = cos x

math 131, techniques of integration

trig integrals

So we obtain
Z

sinn

sinn x dx =

x cos x
n 1
+
n
n

sinn

x dx.

This is what is known as a reduction formula, and it can be used repeatedly to determine integrals of high powers of sin x.
EXAMPLE 8.4.3. Determine

sin5 x dx.

SOLUTION. Use the reduction formula with n = 5.


Z
Z
4

4
sin x cos x
+
5
5

sin5 x dx =

Now use it again with n = 3.


Z

sin x dx =

sin4 x cos x
4
+
5
5

"

sin3 x dx.

sin2 x cos x
2
+
3
3

sin x dx .

Now we can finish the integration in the usual way.


Z

sin5 x dx =

sin4 x cos x
5

4 sin2 x cos x
15

8 cos x
+ c.
15

Reduction Formulas for Large Powers. There are reduction formulas for the other
trig functions as well. They are verified using integration by parts. The most important are (these do not need to be memorized). Repeated application may be
necessary.
R
R
R
R

cosn u du =
sinn u du =

1
n

cosn
1
n

sinn

u sin u +
1

n 1
n

u cos u +

tann u du =

1
n 1

tann

secn u du =

1
n 1

secn

u tan u +

EXAMPLE 8.4.4. Determine

n 1
n

tann

n 2
n 1

tan4 x dx.

tan x
3
tan3 x
=
3

tan3 x
3

sinn

u du
2

u du

u du

SOLUTION. Use the reduction formula with n = 4.


Z
Z
3

tan4 x dx =

cosn

secn

u du

tan2 x dx

tan x
1

1 dx

tan x + x + c.

Notice that we used the fact that tan0 x = 1.

Integrating Products of Powers of Sines and Cosines


We will now develop some guidelines for integrals of the form
Z

sinm x cosn x dx

where either m or n is a positive integer. Notice that both trig functions have the
same argument. The goal is to use the power rule, as in the following: Determine
R
sin x cos4 x dx. You should recognize this as a simple substitution problem; let
u = cos x and du = sin x dx. Then
Z

u5
cos5 x
+c =
+ c.
5
5
Thus, the protocol is to use trig identities and u-substitution to turn the integral
into a simple power rule problem.
sin x cos4 x dx =

u4 du =

10

math 131, techniques of integration

trig integrals

Guidelines for Products of Powers of Sines and Cosines: These general principles can
R
help you solve integrals of the form sinm x cosn x dx.
1. If the power of sine is odd and positive, split off a factor of sine for du and
convert the rest to cosines, let u = cos x, and then integrate. For example,
Z

m=2k +1 odd

use for du
z }| {
Z
z }| {
2k +1
n
2
k
n
sin
x cos x dx = (sin x ) x cos x sin x dx

convert to cosines

z
(1

}|
{
cos2 x )k cosn x sin x dx =

(1

u2 )k un du.

2. If the power of cosine is odd and positive (and the power of sine is even), split
off a factor of cosine for du and convert the rest to sines, let u = sin x, and then
integrate. For example,
Z

n=2k+1 odd

use for du
z }| {
Z
z }| {
2k +1
m
2
k
sin x cos
x dx = sin x (cos x ) cos x dx
m

convert to sines

z
sin x (1
m

}|
{
Z
sin2 x )k cos x dx = um (1

u2 )k du.

3. If both powers of sine and cosine are even and non-negative, make repeated use
1
1
1
2
of the identities sin2 u = 12
2 cos 2u and cos u = 2 + 2 cos 2u to powers of
cosines. Then use reduction formula #1.
4. Use a table of integrals or WolframAlpha or other software. Certainly you should
use this tool in later courses whether in math or other departments.
EXAMPLE 8.4.5 (Using Rule 1). Determine

sin3 x cos2 x dx.

SOLUTION. Since the power of the sine function is odd, we use Guideline #1.

sin3 x cos2 x dx =

=
=
=
=
=

Z
Z

sin2 x cos2 x sin x dx


cos2 x ) cos2 x sin x dx

(1
Z
Z

split off a power of sin x

(1

u2 )u2 du

use a trig id
substitute u = cos x, du =

u2 + u4 du

sin x
expand

u3
u5
+
+c
3
5
cos3 x
cos5 x
+
+c
3
5

EXAMPLE 8.4.6 (Using Rule 2). Determine

sin4 3x cos3 3x dx.

SOLUTION. Both functions have the same argument (angle). The sine function has an

11

math 131, techniques of integration

trig integrals

even power and the power of the cosine function is odd; we use rule #2.
Z

sin4 3x cos3 3x dx =

Z
Z

sin4 3x cos2 3x cos 3x dx


sin4 3x (1

split off a power of cos 3x

sin2 3x ) cos 3x dx

use a trig id

1
u4 (1 u2 ) du
3
Z
1
=
u4 u6 du
3
u5
u7
=
+c
15
21
sin5 3x
sin7 3x
=
+c
15
21

u = sin 3x, 13 du = cos 3x dx


expand

Gosh, be careful of the mental adjustment required for the substitution.


EXAMPLE 8.4.7 (Using Rule 3). Determine

sin2 5x cos2 5x dx.

SOLUTION. The powers of the sine and cosine function are even, we use Guideline #3.

sin2 5x cos2 5x dx =

=
=
=

Z
Z
Z
Z

= 18 x

1
2

1
2

1
4

1
4

1
4

1
4

1
8

1
8
1
160

cos 10x

1
2

+ 12 cos 10x dx

cos2 10x dx
h
i
1
1
2 + 2 cos 20x dx

half angle formula


expand
half angle formula

cos 20x dx

combine terms

sin 20x + c

mental adjustment

Again not too bad! Just be careful.


EXAMPLE 8.4.8 (Using Rule 2). Determine

cos5 x dx.

SOLUTION. Since the sine function does not appear and the power of the cosine

function is odd, we can use Guideline #2. A preferred (?) method might be to use a
reduction formula.
Z

cos5 x dx =

=
=
=

Z
Z
Z
Z

(cos2 x )2 cos x dx
(1

sin2 x )2 cos x dx

(1

u2 )2 du

2u2 + u4 du

split off a power of cos x


use a trig id
u = sin x, du = cos x dx
expand

2u3
u5
+
+c
3
5
2 sin3 x
sin5 x
= sin x
+
+c
3
5

=u

Not too bad!

Integrating Products of Powers of Secant and Tangent


Heres some material from your text that we will use to integrate powers of tangent and secant. You do not need to memorize this. But you do need to know how
to use this information. The idea is very similar to what we used for products of
sines and cosines.

12

math 131, techniques of integration

trig integrals

Guidelines for Products of Powers of Tangents and Secants: These general principles
R
can help you solve integrals of the form tanm x secn x dx.
1. If the power of secant is even and positive, split off sec2 x to use for du and
convert the rest to tangents, then let u = tan x, and integrate. For example,
Z

n=2k even

use for du

z }| {
Z
tan x sec2k x dx = tanm x (sec2 x )k

z }| {
sec2 x dx

convert to tangents

z
}|
{
Z
tanm x (1 + tan2 x )k 1 sec2 x dx = um (1 + u2 )k

du.

2. If the power of tangent is odd and positive (and the power of secant is odd),
split off sec x tan x for du and convert the rest to secants, let u = sec x, and then
integrate. For example,
Z

m=2k +1 odd

z }| {
Z
tan2k+1 x secn x dx = (tan2 x )k secn

use for du

convert to secants

z
}|
{
x sec x tan x dx

z
}|
{
(sec2 x 1)k secn

x sec x tan x dx =

( u2

1) k u n

3. If m is even and n is odd, convert the tangents to secants and use the reduction
formula for powers of the secant:
Z

m=2k even n odd

convert to secants

z }| { z }| {
}|
{
Z z
tan2k x secn x dx =
(sec2 x 1)k secn x dx.

4. In real life, use WolframAlpha, or look in a table of integrals.


EXAMPLE 8.4.9 (Using Guideline 2). Determine

tan3 x sec3 x dx.

SOLUTION. The powers of both secant and tangent are odd, so Guideline #2 above

applies.
Z

tan3 x sec3 x dx =

=
=
=

Z
Z
Z
Z

tan2 x sec2 x (sec x tan x ) dx

(sec2 x

split off a secant-tangent

1) sec2 x (sec x tan x ) dx

( u2

1)u2 du

u4

u2 du

convert to secants
u = sec x,

expand

u5
u3
+c
5
3
= 15 sec5 x 13 sec3 x + c

EXAMPLE 8.4.10 (Using Guideline 1). Determine

du = sec x tan x

tan4 x sec4 x dx.

du.

13

math 131, techniques of integration

trig integrals

SOLUTION. The powers of both secant and tangent are even, so Guideline #1 applies.
Z
Z

tan4 x sec4 x dx =

=
=
=

Z
Z
Z

tan4 x sec2 x sec2 x dx

split off a secant-squared

tan4 x (1 + tan2 x ) sec2 x dx

convert to tangents

u4 (1 + u2 ) du

u = tan x,

u4 + u6 du

du = sec2 x
expand

u5
u7
+
+c
5
7
= 15 tan5 x + 17 tan7 x + c

EXAMPLE 8.4.11 (Using Guideline 3). Determine

tan2 x sec x dx.

SOLUTION. The powers of tangent is even and the power of secant is odd, so Guide-

line #3 applies.
Z

tan2 x sec x dx =

Z
Z

(sec2 x
sec3 x dx

sec x tan x
2
sec x tan x
=
2
sec x tan x
=
2

1) sec x dx
Z

convert to secants

sec x dx

1
2

sec x dx

expand
Z

sec x dx

1
sec x dx
2
1
ln | sec x + tan x | + c.
2

reduction formula
combine, simplify

Take-Home Message. OK, the goal here is to be able to use these guidelines for
products of trig functions. Heres what you must be able to do without looking
back to the guidelines:
1. You need to KNOW the four key identities.
2. You need to be able to integrate low powers (n = 1 or 2) of the four main trig
functions.
3. You need to be able to integrate products of the form sinn x cosm x when at
least one of the powers is an odd positive integer by splitting off an appropriate
factor, using a key trig id, and then using u-substitution (Guidelines 1 and 2 for
Products of Sines and Cosines).
4. You do NOT need to memorize the reduction formul, do not need to memorize the guidelines for products of the form tanm x secn x.

14

Integration by Triangle Substitutions


The Area of a Circle
So far we have used the technique of u-substitution (i.e., reversing the chain rule)
and integration by parts (reversing the product rule) to extend the list" of functions that we can antidifferentiate. Remember that we are in the antidifferentiation
business because the Fundamental Theorem (FTC) says that if F is an antiderivaRb
tive of f , then the area under f on the interval [ a, b] is a f ( x ) dx = F (b) F ( a).
This theorem solves" the area problem, at least for those functions whose antiderivatives we know. But even with our new integration techniques there are
many integrals we cannot yet do, such as:
EXAMPLE 7.2.1. Find the area of the semi-circle of radius r. Since middle school we have
been told that the area of a circle is r2 , so the area of a semi-circle is 12 r2 . But why is
this formula valid? Recall that the entire circle of radius r centered at the origin is the set
of points that satisfiesx2 + y2 = r2 . It follows that the upper semi-circle is given by the
function y = f ( x ) = r2 x2 on the interval [r, r ]. So the area we seek is

Z r p
r

r2 x2 dx.

Its scandalous but true: none of thetechniques weve developed so far will help us find
an appropriate antiderivative. Still r2 x2 should remind us of right triangles, and weve
excised one such triangle from the figure below.
................................
...........
........
........
........
......
.... ........
.....
.
.
.
.
. .. ........
....
.
.
.
.
...
..
.. ..
.
.
...
.
...
...
..
...
...
..
...
...
..
.
....
.
...
.
...
...
...
.
.
...
.
...
...
.
..

f (x) =

r2

..
.....
... ..
... ...
... ...
.
.
....
...
...
...
...
...
...
.
...
.
.
.
..
................................

x2

r2

x2

The sides of the triangle are related to each other via the trigonometric functions of
the angle . We can use these to set up a fancy u-substitution, or more appropriately, a
-substitution.

p
r2 x2
= sin
r2 x2 = r sin
r
x
= cos
x = r cos
r

dx = r sin d

Notice that each part of the original integrand can now be written in terms of the angle .
Of course, we need to write the x-limits of integration as -limits. When
x = r = r cos
x = r = r cos

cos = 1
cos = 1

=
=0

This makes sense: changes from to 0 as you move around the semi-circle clockwise

Figure 7.1: Left:


A semi-circle with
equation y = r2 x2 . Right: A
corresponding right triangle.

math 131, techniques of integration iv

triangle substitutions

from r to r. Now the integral can be rewritten in terms of and solved.


Z r p
r

r2 x2 dx =

Z 0

r sin (r sin ) d = r2

Z 0

sin2 d = r2

Z 0

1
2

12 cos(2 ) d

 0

14 sin(2 )

h

i
= r 2 (0 0)
0
2
r2
=
2

= r 2

1
2

From this it follows that the area of a circle of radius r is r2 .

7.3 General Triangle Substitutions


In general, triangle substitutions are based on, well, triangles! Right triangles, in particular.
Three different substitutions arise depending on the integrand and how the sides of the
triangle are labeled. The triangles are based on the corresponding forms of the square
roots

2 + a2 ,
that
arise
from
the
Pythagorean
theorem
in
these
triangles.
The
three
cases
are:
x

x2 a2 , and a2 x2 .

Case 1: Integrals involving x2 + a2

In this case, a2 + x2 must correspond to the hypotenuse of the right triangle. (Why?)
So with the legs of the triangle labeled as below, we have:

x2

...
.....
... ..
.. ....
.
.
. .
... ...
...
...
2 .....
...
.
.
...
...
...
.
.
.
.
...
...
...
.
.
.
.
...
..
.
.
.
.........................................

+a

x = a tan
x

dx = a sec2 d

x2 + a2 = a sec

a
EXAMPLE 7.3.1. Calculate the following indefinite integral:

dx.
x2 x2 + 4

Using the triangle substitution in the box above with a2 = 4 or a = 2, we have:


Z

x2 x2 + 4

dx =

1
2 sec2 d =
2
4 tan 2 sec

1
4

sec
d =
tan2

1
4

1
4

1
4

1
cos
sin2
cos2

cos
d
2
sin
Z
(sin )2 cos d

= 41 (sin )1 + c
We must convert our answer back to a function of x. Look back at the triangle above. Notice
that sin = x2 . So
x +4

14 (sin )1 + c = 14

 1

x2 + 4

x2 + 4
+ c.
4x

One question you may have is how should you select which leg of the triangle should
correspond to x and which to a? The answer is that the substitution will work no matter
which you use, but it will be easier with x as the opposite side to angle . If we had let x be
the adjacent side, then x = a cot , a less familiar trig function.
EXAMPLE 7.3.2. Calculate the indefinite integral:

1
16 + x2

dx.

math 131, techniques of integration iv

triangle substitutions

Using the triangle substitution above with a2 = 16 or a = 4, we have:

4 sec2
2
d = sec d = ln | sec + tan | + c = ln | 164+ x + 4x | + c
4 sec
16 + x2

Case 2: Integrals involving x2 a2

dx =

In this case, x must correspond to the hypotenuse of the right triangle.


(Why?) Again
we have our choice of how to label the legs, one side a and the other x2 a2 . With the
selection made below, x = a sec . What would x equal if we had let a be the side opposite
?
....
... ...
... ..
.. ....
.
.
.. ...
...
....
...
...
...
...
...
.
.
.
.
...
...
.
....
.
.
..
.
...
...
.
.
.
.
........................................

x = a sec
x2

a2

dx = a sec tan d

x2 a2 = a tan

a
EXAMPLE 7.3.3. Evaluate the definite integral:

Z 2 2
x 1

dx.

Use the triangle substitution above with a2 = 1 or a = 1. We also have to change limits.
x = 1 = sec
x = 2 = sec

=0
cos =

1
2

= /3

Now proceed with the substitution:


Z 2 2
Z /3
Z /3
Z /3
tan
x 1
tan2 d =
sec2 1 d
dx =
sec tan d =
x
sec
1
0
0
0
/3

= tan
0

= 3 /3

Case 3: Integrals involving

a2 x 2

This was situation in the semi-circle example. In this case, a must correspond to the
hypotenuse of
the right triangle. (Why?) Again we have our choice of how to label the
legs, one side a2 x2 and the other x. With the selection made below, x = a sin . In the
circle example we chose to label the legs in the other way because of the geometry involved.
Obviously it worked out, but it required us to carry along a minus sign throughout the
problem. The choice below is usually simpler.
...
.....
... ..
.. ....
.
.
.
... ....
...
...
...
...
..
.
.
...
...
.
...
.
.
.
...
...
.
....
.
.
..
.
..
.
.
.
.........................................

x = a sin
x

dx = a cos d

a2 x2 = a cos

a2 x 2

EXAMPLE 7.3.4. Calculate the following indefinite integral:

x2
25 x2

dx.

Use the substitution above with a2 = 25 or a = 5. We use a reduction formula (see the
Appendix to these notes) to do the integral.
Z

x2
25 x2

dx =

25 sin2
5 cos d = 25
5 cos

1
2

25
2

sin2 d = 25

12 cos(2 ) d
25
4

sin(2 ) + c

math 131, techniques of integration iv

triangle substitutions

To solve for , look back at the original triangle where

x
5

= sin arcsin

x
5

= and

25 x2
5

= cos . The simplest way to finish the problem is to make use of a Double Angle
Formula:
sin(2 ) = 2 sin cos .
Then
25
2

25
4

sin(2 ) + c =

25
2

25
4 (2 sin cos ) + c

x
5

25
2

x
5

25
2

25
x
1
arcsin x
2
5
2

arcsin

25 x2
5

+c

25 x2 + c.

7.4 Additional Examples


In this last section we extend the triangle substitution idea to integrals that at first dont
appear to have anything to do with triangles because a square root does not immediately
appear in them.
1
dx.
( x2 + 6)3/2
Well there is sort of
a square root lurking in the background here. In this case, ( x2 + 6)3/2
may be thought of as ( x2 + 6)3 . So the appropriate triangle is:

EXAMPLE 7.4.1. Calculate the following indefinite integral:

.
.....
... ..
.. ....
.
.
. .
... ...
...
...
...
...
2
..
.
...
.
...
...
.
.
.
...
.
...
...
.
.
.
.
...
..
.
..
.
..........................................

x +6

x=

6 tan

dx = 6 sec2 d

x2 + 6 = 6 sec

Notice that one of the sides of the triangle is 6; we may not always have perfect squares.
Using the triangle substitution in the box we have:
Z
Z
Z
Z
Z
1
6 sec2
6 sec2
1
1
1

d
=
d
=
d
=
cos d
dx
=
6
6
sec
( x2 + 6)3/2
( 6 sec )3
6 6 sec3
=

1
6

sin + c
x
=
+ c.
6 x2 + 6

EXAMPLE 7.4.2. Determine the following indefinite integral:

4
dx.
4 x2

This time we have to think of 4 x2 as ( 4 x2 )2 to use a triangle.


...
.....
... ..
.. ....
.
.
.
... ....
...
...
...
...
..
.
.
...
...
.
...
.
.
.
...
...
.
...
.
.
.
...
..
.
.
.
........................................

x = 2 sin
x

dx = 2 cos d

4 x2 = 2 cos

4 x2

Now we can rewrite the integral.


Z

4
dx =
4 x2

4
2 cos d =
(2 cos )2

8 cos
d = 2
4 cos2

=2

Z
Z

sec d
ln | sec( ) + tan | d

2
+ x 2|+c
4 x 2
4 x
2+ x

2 ln |
|+c
4 x 2

= 2 ln |
=

x)
= ln | (24+
|+c
x2

+x | + c
= ln | 22
x

= ln |2 + x | ln |2 x | + c.

math 131, techniques of integration iv

triangle substitutions

We will see this same integral again, shortly, and it will be solved very differently.
EXAMPLE 7.4.3. Evaluate the definite integral:

Z 1
1/2

x3
4x2 1

dx.

This time the triangle is obvious, but care is required to label the sides correctly.
..
.....
... ...
... ...
.
. .
... ....
..
...
...
...
...
.
...
.
.
.
...
...
.
...
.
.
.
...
...
.
...
.
.
.
..
.........................................

2x

x=
4x2 1

1
2

sec
1
2

dx = sec tan d

4x2 1 = tan

You could do the indefinite integral and convert back to x to avoid changing the limits.
But lets actually change them
x = 12 =
x=1=

1
2 sec
1
2 sec

1 = sec
2 = sec

=0
cos =

1
2

= /3

Now proceed with the substitution (and use the guidelines for integrating powers of sec ):
Z 1

1/2

x3
4x2 1

dx =

Z /3 1
( 2 sec )3
0

tan

12 sec tan d =

1
16

Z /3
0

sec4 d

1 /3
sec2 sec2 d
16 0
Z
1 /3
(1 + tan2 ) sec2 d
=
16 0

u2

du

z }| { z }| {
Z
1 /3
=
sec2 + tan2 sec2 d
16 0


1
tan3 /3
=
tan +

16
3
0
"
#
3

( 3)
3
1
3+
0 =
=
16
3
8

7.5

Problems

1. Try these problems. A variety of techniques are required, not just triangle substitutions.
Z
Z
Z 2
x 9
1
1

dx
(b)
dx
(c)
dx
(a)
2
3/2
x
(1 + x )
16 x2
Z 2
Z
Z p
1
x2

dx
(e)
(d)
dx
(f )
x 1 x2 dx
0
4 x2
x x2 9
Z
Z
Z
4x
4
4
dx
(h)
dx
(i)
dx
(g)
4 + x2
4 + x2
4 x2
Z
Z
Z
5
25 + x2
x
1

(j)
dx
(l)
dx
dx
(k)
4
x
0
25 + x2
25 x2
(m)
(o)

Z
Z 0
5

x3
x2

dx

25
p
25 x2 dx

dx
25 + x2
Z
1

(p)
dx
x2 x2 25
(n)

2. Find the arc length of the parabola f ( x ) = x2 on the interval [0, 1]. You will have to use a
trig substitution. Make sure that you switch the limits of the integration.

math 131, techniques of integration iv

Answers
1. Caution. . . , its easy to have made a typo in these answers. Remember: +c
 2 
p
x
x 9
(a)
(b)
x2 9 3 arctan
3
1 + x2
(c) arcsin( x/4)
(d) 2 1
 2 
(1 x2 )3/2
x 9
(f )
(e) 13 arctan
3
3
(g) 2 arctan( x/2)
(h) 2 ln |4 + x2 |
!3



2
2+x
1
25
+
x

(i) 2 ln
(j)

75
x
4 x2

(k) 5( 2 1)
(l) arcsin( x/5)


25 + x2
2
3/2
p
x
(25 x )

2
(m) 25 25 x +
(n) ln
+

3
5
5

x2 25
(o) 25/4
(p)
25x

2. Answer:

2+ln | 2+1|)
2

7.6 Appendix: Common Trigonometric Formulas and Antiderivatives


Below are listed several integral formulas for various powers of trig functions.
1. Degree 2 Sine and Cosine Functions. One simple way to do these is to use trig identities.Z Know these.Z
cos2 u du =

(a)

(b)

sin2 u du =

1
2

+ 12 cos 2u du.

1
2

21 cos 2u du.

2. Low Powers of the Tangent and Secant Functions. These are done with simple identities.Z Know these.
Z
sin u
tan u du =
(a)
du = ln | sec u| + c.
cos
u
Z
Z
tan2 u du =

(b)

sec2 u 1 du = tan u u + c.

sec2 u + sec u tan u


du = ln | sec u + tan u| + c.
sec u + tan u
3. Useful Double Angle Formula: sin(2 ) = 2 sin cos .
(c)

sec u du =

4. Reduction Formulas for Large Powers. These are verified using integration by parts.
Repeated application may be necessary.
Z
R
1
cosn2 u du
(a)
cosn u du = n1 cosn1 u sin u + n
n
(b)
(c)
(d)

R
1
sinn u du = n1 sinn1 u cos u + n
sinn2 u du
n

1
n 1 u
tann u du = n
1 tan

R
1
n2 u tan u + n2
secn u du = n
secn2 u du
1 sec
n 1

tann2 u du

5. Degree 2 Sine and Cosine Functions Again. If we apply the reduction formulas as
for the sine and cosine functions when n = 2, we get a different form of the earlier
answer. These new forms are better to use with indefinite integrals involving triangle
substitutions because it is easier to convert back from u to the original variable x. (Know
either these or those in #1.)
Z
R
(a)
cos2 u du = 21 cos u sin u + 21 1 du = 12 cos u sin u + 21 u + c
(b)

sin2 u du = 12 sin u cos u +

1
2

1 du = 12 sin u cos u + 21 u + c

triangle substitutions

Rational Functions and Partial Fractions


Our final integration technique deals with the class of functions known as rational
functions. Recall from Calculus I that
DEFINITION 7.1. A rational function1 is a function that is the ratio of two polynomials

r(x) =

p( x )
,
q( x )

where p( x ) and q( x ) are polynomials. (Remember a polynomial has the form


p ( x ) = a n x n + a n 1 x n 1 + + a 1 x + a 0 ,
where the ai are real constants and n is a non-negative integer and is called the degree of the
polynomial.)

Here are several examples of rational functions; identify the polynomials p( x )


and q( x ).
r(x) =

x2 + 7
3x5 + 7x

s( x ) =

2x3 9x + 1
x

t( x ) =

1
x2 + 1

r ( x ) = 3x 2 + 2x + 1 is also rational because it can be written as a ratio of two


polynomials
2x3 + x2 + 3
3
.
r ( x ) = 2 + 2x + 1 =
x
x2
A couple of non-examples include
x3/2 + 7
because the term x3/2 is not a polynomial since the power is not
3x + 7x
a non-negative integer.

z( x ) =

s( x ) =
tion.

sin( x2 + 1)
is not a polynomial since it contains a trigonometric funcx

Remember from Calculus I that rational functions are continuous and differentiable at all points in their domains, i.e., at all points where the denominator is not
0.
Our main concern in this chapter is to determine when such rational functions
can be integrated. From our earlier work, we know that any interval on which a
rational function is continuous it is also integrable. But, how do we actually find
an antiderivative for the function? For instance, to check that you can integrate

Here rational means ratio, as in the


ratio of two polynomials.
1

math 131, techniques of integration v

partial fractions

3x2 1
1
dx. But what about very simdx, and
1
+
x
xZ3 x
Z
2x + 1
4
ilar looking functions such as
dx? Why are these integrals
dx or
x3 x
4 x2
not so easy to do? Over the next few sections, we examine a series of special cases
of rational functions that we will see are relatively easy to integrate using a technique known as partial fractions.
Z

1
dx,
1 + x2

x
dx,
1 + x2

7.1 Special Cases: Linear Factors with Degree p( x ) < q( x )


There are several techniques that can be used integrate rational functions. We will
concentrate on a single technique and a couple of simple variations that work with
rational functions of a particular type.
Assume we have a rational function of the form
r(x) =

p( x )
,
q( x )

where degree of p( x ) < degree of q( x ) and q( x ) can be factored into linear factors
(factors of degree 1). Some examples include

2x + 1
2x + 1
2x + 1
=
=
where all the factors in the denominator
x ( x 1)( x + 1)
x3 x
x ( x 2 1)
are linear and different and degree p( x ) = 1 and degree q( x ) = 3.
4
4
=
; the two factors in the denominator are linear and
(2 x )(2 + x )
4 x2
different and degree of p( x ) < degree of q( x ).
6x 2
6x 2
=
where all the factors in the denominator are linear
( x 3)( x + 1)
x2 2x 3
and different and degree of p( x ) < degree of q( x ).
x+4
x+4
=
; the three factors in the denominator are linear and
x x ( x + 1)
x3 + x2
degree of p( x ) < degree of q( x ).

On the other hand the rational function


2x + 1
2x + 1
=
x3 + x
x ( x 2 + 1)
does not satisfy our criterion above because the denominator contains a factor of
degree 2, in particular, x2 + 1 cannot be written as a product of linear factors.
We will use a technique called partial fractions to integrate the special rational
functions above.2

The Key to Partial Fractions


Lets look at a couple of ordinary fractions and how they can be rewritten in simpler terms. Notice
1 1
1
=
20
4 5
or
10
1 1
= + .
21
3 7

Partial fractions can be used to integrate other types of rational functions.


Your text has further examples.
2

math 131, techniques of integration v

partial fractions

In each case the original fraction has been rewritten in terms of simpler component
fractions. The idea is to do the same for rational functions. For example, can we
write
4
A
B
as
+
2x 2+x
4 x2
where 2 x and 2 + x are the linear factors of 4 x2 ? If we could, then putting the
two simpler pieces back over a common denominator would give us
A
( A B) x + (2A + 2B)
4
B
2A + Ax + 2B Bx
=
=
.
+
=
2
2
2x 2+x
4x
4x
4 x2

(7.1)

The first and last rational functions in (7.1) are equal and have the same denominator. The only way that this can happen is if the numerators are also the same. So
(7.1) means
4 = ( A B) x + (2A + 2B).
(7.2)
But (7.2) is true if and only if the x-terms on each side are equal and the constants
on each side are equal. Since there are no xs on the left side and A B on the
right, we must have
xs : 0 = A B.
(7.3)
Comparing the constant terms in the same way we have
constants :

4 = 2A + 2B.

(7.4)

From (7.3) we see that A = B and using this in (7.4) gives


4 = 4A.
Thus, A = 1. Putting A = 1 in (7.3) or (7.4) makes B = 1. So we see that
4
1
1
=
+
.
2
2x 2+x
4x

(7.5)

Check that this is correct! We describe this process by saying that we have rewritten 44x2 as 21 x + 2+1 x using partial fractions.
Z
4
So what! How do we use this? Well, suppose we need to solve
dx.
4 x2
Using (7.5) we can rewrite it and then integrate (using a mental adjustment):

Z
Z 
4
1
1
+
dx
dx =
2x 2+x
4 x2


2 x

+ c.
= ln |2 x | + ln |2 + x | + c = ln
2 + x

7.2

The Easiest Case: Distinct Linear Factors

We can always carry out the same sort of process as above under the following
circumstances:3
r(x)
be a rational function. Assume that the denominator
q( x )
q( x ) of the rational function factors into distinct linear and the degree of the numerp( x )
ator p( x ) is less than the degree of the denominator q( x ). Then r ( x ) =
can be
q( x )
rewritten using partial fractions.
The Easiest Case: Let r ( x ) =

This is not so easy to prove in general;


usually one sees the proof of this result
in a graduate-level abstract algebra
course.
3

math 131, techniques of integration v

partial fractions

EXAMPLE 7.2.1 (Partial Fractions: Easiest Case). Heres another example that uses partial frac-

tions. Determine

3
dx.
x2 + 3x + 2

SOLUTION. This is not an integral that we can immediately do, even with integration
by parts. 4 So we try partial fractions. Notice that the degree of the numerator is less
than the degree of the denominator (0 < 2) and the denominator factors into distinct
A
B
linear factors: ( x + 1)( x + 2). We form the partial fractions x+
1 and x +2 , where A and
B are constants. We solve for A and B as we did in the previous example.

3
A
B
3
=
=
+
.
( x + 1)( x + 2)
x+1
x+2
x2 + 3x + 2
By putting the last terms together again, we get
Ax + 2A + Bx + B
( A + B) x + (2A + B)
3
=
=
.
x2 + 3x + 2
x2 + 3x + 2
x2 + 3x + 2
Since the denominators are the same, the numerators must be the same, too. In particular there are as many xs on the left side as on the right. There are none on the left
and A + B on the right side. Similarly for the constants.
xs:
constants:

0
3

=
=

A+B
2A + B

Subtracting the first equation from the second gives


3 = A.
Substituting A = 3 into first or second equation makes B = 3. So we see that
3
3
3
=

.
x+1
x+2
x2 + 3x + 2
Now on to integration:
Z

3
dx =
x2 + 3x + 2

Z 

3
3

x+1
x+2


dx



x + 1

+ c.
= 3 ln | x + 1| 3 ln | x + 2| + c = 3 ln
x + 2
EXAMPLE 7.2.2 (Partial Fractions: Easiest Case). Determine
Z

6x 2
dx.
x2 2x 3

SOLUTION. This is not an integral that we can immediately do with substitution or


integration by parts. So we try partial fractions. The degree of the numerator is less
than the degree of the denominator (1 < 2) and the denominator factors into distinct
A
B
linear factors: ( x 3)( x + 1). We form the partial fractions x
3 and x +1 , where A and
B are constants.

6x 2
6x 2
A
B
Ax + A + Bx 3B
=
=
+
=
.
( x 3)( x + 1)
x3
x+1
x2 2x 3
x2 2x 3
So comparing the numerators of the first and last functions
xs:
constants:

6
2

=
=

A+B
A 3B

Subtracting the second equation from the first gives


8 = 4B.
So B = 2. Using this in the first or second equation makes A = 4. Now on to integration (using an adjustment):

Z
Z 
6x 2
4
2
dx
=
+
dx = 4 ln | x 3| + 2 ln | x + 1| + c.
x3
x+1
x2 2x 3

4
Dont confuse the terms partial fractions
with integration by parts.

math 131, techniques of integration v

partial fractions

YOU TRY IT 7.1. Does it matter which letters you put over each linear factor? What would
the numerator in the original integral have to be to make the problem a substitution problem?
EXAMPLE 7.2.3 (Partial Fractions: Easiest Case). Heres another quick example. Determine
Z

2x + 5
dx.
x2 + 2x 8

SOLUTION. Notice this is not quite a substitution integral, but partial fractions will

work. The degree of the numerator is less than the degree of the denominator (1 < 2)
and the denominator factors into distinct linear factors: ( x 2)( x + 4).
2x + 5
2x + 5
A
B
Ax + 4A + Bx 2B
=
.
=
+
=
( x 2)( x + 4)
x2
x+4
x2 + 2x 8
x2 + 2x 8
Comparing the numerators of the first and last functions and solving for A and B
gives
xs:
2 = A+B
4 = 2A + 2B
constants: 5 = 4A 2B
5 = 4A 2B

9
So A =
Z

x2

3
2

and B =

1
2.

The integration becomes:


!
Z
3
2

2x + 5
dx =
+ 2x 8

x2

1
2

x+4

dx =

3
2

6A

ln | x 2| +

1
2

ln | x + 4| + c.

YOU TRY IT 7.2. What would the numerator in the original integral have to be to make the
problem a substitution problem?

7.3 A Complication: Higher Degrees


The first complication that arises is is that the denominator of the rational function
may factor into three or more distinct linear factors. The solution method works
the same way as above, but it may be more complicated to find the constants.
EXAMPLE 7.3.1 (Partial Fractions: Three Distinct Linear Factors). Determine

2x2 6x + 2
dx.
x3 3x2 + 2x

SOLUTION. Check that this is not quite a substitution integral. However, the degree

of the numerator is less than the degree of the denominator (2 < 3) and the denominator factors into distinct three linear factors: x ( x2 3x + 2) = x ( x 1)( x 2). We
form the partial fractions with a different constant for each linear factor:
2x2 6x + 2
A
B
C
2x2 6x + 2
=
= +
+
3
2
x ( x 1)( x 2)
x
x1
x2
x 3x + 2x
A( x 1)( x 2) + Bx ( x 2) + Cx ( x 1)
=
x3 3x2 + 2x
2
Ax 3Ax + 2A + Bx2 2B + Cx2 Cx
=
.
x3 3x2 + 2x
Compare like terms in the numerators of the first and last functions.
x2 s:
xs:
const:

2
6

=
=

A+B+C
3A 2B C

2A

A=1

1
3

=
=

B+C
2B C

Notice that we used the value A = 1 to simplify the first two equations. It follows that
B = 2 and C = 1. The integration becomes:
Z

2x2 6x + 2
dx =
x3 3x2 + 2x

1
2
1
+

dx = ln | x | + 2 ln | x 1| ln | x 2| + c.
x
x1
x2

Answers to you try it 7.1 : No. It


would have to be a multiple of 2x 2.

math 131, techniques of integration v

partial fractions

YOU TRY IT 7.3. What would the numerator in the original integral have to be to make the
problem a substitution problem?
EXAMPLE 7.3.2 (Partial Fractions: Three Distinct Linear Factors). Determine

x2 + 4x 1
dx.
x3 x

SOLUTION. Check that this is not a substitution integral. However, the degree of the
numerator is less than the degree of the denominator (2 < 3) and the denominator
factors into distinct three linear factors: x ( x2 1) = x ( x 1)( x + 1). We form the
partial fractions with a different constant for each linear factor:

A
B
C
A( x2 1) + Bx ( x + 1) + Cx ( x 1)
x2 + 4x 1
= +
+
=
3
x
x1
x+1
x x
x3 x
2
2
Ax A + Bx + B + Cx2 Cx
.
=
x3 x
Compare like terms in the numerators of the first and last functions.
x2 s:
xs:
const:

1
4

=
=

A+B+C
BC

A=1

0
4

=
=

B+C
BC

2B

Notice that we used the value A = 1 to simplify the first two equations. It follows that
B = 2 and C = 2. The integration becomes:
Z

x2 + 4x 1
dx =
x3 x

1
2
2
+

dx = ln | x | + 2 ln | x 1| 2 ln | x + 1| + c
x
x1
x+1


x 1
+ c.
= ln | x | + 2 ln
x + 1

EXAMPLE 7.3.3 (Partial Fractions: Three Distinct Linear Factors). Determine

4x + 28
dx.
( x + 1)( x2 4x + 3)

SOLUTION. Check that this is not a substitution integral. However, the degree of the
numerator is less than the degree of the denominator (1 < 3) and the denominator
factors into distinct three linear factors: ( x + 1)( x2 4x + 3) = ( x + 1)( x 1)( x 3).
We form the partial fractions with a different constant for each linear factor:

4x + 28
A
B
C
A( x2 4x + 3) + B( x2 2x 3) + C ( x2 1)
=
+
+
=
2
x+1
x1
x3
( x + 1)( x 4x + 3)
( x + 1)( x2 4x + 3)

( A + B + C ) x2 + (4A 2B) x + 3A 3B C
.
( x + 1)( x2 4x + 3)

Compare like terms in the numerators of the first and last functions.
x2 s:
xs:
const:
Add all:

0
4

=
=

28
32

=
=

A+B+C
4A 2B
3A 3B C
4B

C=5

A=3

B = 8

The integration becomes:


Z

4x + 28
dx =
( x + 1)( x2 4x + 3)

3
8
5

+
dx
x+1
x1
x3

= 3 ln | x 1| 8 ln | x + 1| + 5 ln | x 3| + c
YOU TRY IT 7.4. What would the numerator in the original integral have to be to make the
problem a substitution problem?

math 131, techniques of integration v

7.4

partial fractions

A Second Complication: Repeated Factors

Another complication that can arise is that the denominator of the rational function factors into linear factors, but the factors are not distinct. That is, there are
repeated linear factors. The solution in this situation is to include a term for every
power of every linear factor that divides the denominator.
EXAMPLE 7.4.1 (Partial Fractions: Repeated Linear Factors). Determine

3x2 7x + 2
dx.
x3 2x2 + x

SOLUTION. This is not a substitution integral. However, the degree of the numerator

is less than the degree of the denominator (2 < 3) and the denominator factors into
repeated linear factors:
x ( x2 2x + 1) = x ( x 1)( x 1) = x ( x 1)2 .
We form the partial fractions with a different constant for each power of each factor
that divides the denominator. Note: There are terms for both x 1 and ( x 1)2 .
Warning! Look very carefully at the numerator in the step where the common denominator is formed.
3x2 7x + 2
A
A( x 1)2 + Bx ( x 1) + Cx
B
C
=
=
+
+
x
x1
x ( x 1)2
( x 1)2
x ( x 1)2

Ax2 2Ax + A + Bx2 Bx + Cx


.
x ( x 1)2

Compare like terms in the numerators of the first and last functions.
x2 s:
xs:
const:

3
7
2

=
=
=

A+B
2A B + C
A

1=B
C = 2
A=2

Notice that we used the value A = 2 to simplify the first two equations. The integration becomes:
Z

3x2 7x + 2
dx =
x3 2x2 + x

2
1
2
+

dx
x
x1
( x 1)2

= 2 ln | x | + ln | x 1| + 2( x 1)1 + c.
Be careful! The final integral requires the power rule and a mini-substitution.
YOU TRY IT 7.5. What would the numerator in the original integral have to be to make the
problem a substitution problem?
EXAMPLE 7.4.2 (Partial Fractions: Repeated Linear Factors). Determine

3x2 2x 3
dx.
x3 x2

SOLUTION. This is not quite a substitution integral. However, the degree of the nu-

merator is less than the degree of the denominator (2 < 3) and the denominator
factors into repeated linear factors: x2 ( x 1). We form the partial fractions with a
different constant for each power of each factor that divides the denominatorthere
are terms for both x and x2 . Look very carefully at the numerator in the step where
the common denominator is formed.
3x2 2x 3
A
C
B
Ax ( x 1) + B( x 1) + Cx2
=
+
+
=
x
x1
x 2 ( x 1)
x2
x 2 ( x 1)

Ax2 Ax + Bx B + Cx2
.
x 2 ( x 1)

math 131, techniques of integration v

partial fractions

Compare like terms in the numerators of the first and last functions.
x2 s:
xs:
const:

=
=
=

3
2
3

C = 2

A+C
A + B
B

A=5
B=3

Notice that we used the value B = 3 to simplify the first two equations. The integration becomes:
Z

3x2 2x 3
dx =
x3 x2

5
2
3
+ 2
dx = 5 ln | x | 3x 1 2 ln | x 1| + c.
x
x1
x

YOU TRY IT 7.6. What would the numerator in the original integral have to be to make the
problem a substitution problem?
YOU TRY IT 7.7. Determine

Answer to you try it 7.7 : 2 ln | x + 1|


3
2 ln | x 1| x
1.

7x
dx.
( x + 1)( x 1)2

EXAMPLE 7.4.3 (Partial Fractions: Repeated Linear Factors). Heres a different one: Determine

x2
dx.
( x + 1)3

SOLUTION. The degree of the numerator is less than the degree of the denominator
(2 < 3) and the denominator factors into repeated linear factors: ( x + 1)3 = ( x +
1)( x + 1)( x + 1). We form the partial fractions with a different constant for each
power of each factor that divides the denominator: ( x + 1), ( x + 1)2 , and ( x + 1)3 .
Look very carefully at the numerator in the step where the common denominator is
formed.

x2
A
C
A ( x + 1)2 + B ( x + 1) + C
B
=
+
=
+
3
2
3
x+1
( x + 1)
( x + 1)
( x + 1)
( x + 1)3

Ax2 + 2Ax + A + Bx + B + C
.
( x + 1)3

Compare like terms in the numerators of the first and last functions.
x2 s:
xs:
const:

=
=
=

1
0
0

A=1

A
2A + B
A+B+C

B = 2
C=1

Notice that we used the value A = 1 to simplify the last two equations. The integration becomes:
Z

x2
dx =
( x + 1)3

1
2
1

+
dx
x+1
( x + 1)2 ( x + 1)3

= ln | x + 1| + 2( x + 1)1 21 ( x + 1)2 + c.
Alternative Solution. We could have used a u-substitution to solve this problem. Let
u = x + 1. Then du = dx. Since u = x + 1, then x = u 1 so x2 = (u 1)2 =
u2 2u + 1. Therefore
Z

x2
dx =
( x + 1)3

u2 2u + 1
du =
u3

1
2
1
2 + 3 du
u
u
u

= ln |u| 2u1 12 u2 + c
= ln | x + 1| + 2( x + 1)1 21 ( x + 1)2 + c,
just as above. Which method would you have used? Why?

7.5 Problems

math 131, techniques of integration v

partial fractions

1. Try integrating these rational functions (answers below). These are a bit harder than
those in the text. Some have three factors. Others have repeated factors.
(a)

4t2 3t 4
dt
t3 t2 2t

(b)

2x2
dx
( x 1)2 ( x + 1)
Z
2x 1
(h)
dx
( x + 2)2

x
dx
( x 1)( x + 1)( x + 2)
Z
2x
(g)
dx
( x 2)( x2 1)
(d)

t+7
dt
(t + 1)(t2 4t + 3)

(e)

(c)
(f )

x+6
dx
x2 x 6

4x + 4
dx
( x 2)2 x

2. Try these similar looking problems.


(a)

10
dx
25 + x2

(b)

10x
dx
25 + x2

(c)

10
dx
25 x2

3. Have you finished all the ones above? Do these similar looking integrals.
(a)

dx

4 + x2
Z
4x
(c)
dx
(4 + x2 )3/2

(b)
(d)

4
dx
4 x2
p
4 x2 dx

Z 2
2

Answers to Practice Problems


1
1
2
t + t2 + t+1 dt = 2 ln | t | + ln | t 2| + ln | t + 1| + c
3/4
5/4
2
3
5
t+1 + t3 t1 dt = 4 ln | t + 1| + 4 ln | t 3| 2 ln | t 1| + c
4/5
9/5
9
4
x 3 x +2 dt = x + 5 ln | x 3| 5 ln | x + 2| + c
1/6
1/2
2/3
1
1
2
x 1 + x +1 x +2 dx = 6 ln | x 1| + 2 ln | x + 1| 3 ln | x + 2| + c
1/2
3/2
1
3
1 + 1 ln | x + 1| + c
x 1 + ( x 1)2 + x +1 dx = 2 ln | x 1| ( x 1)
2

1. (a)

(b)

(c)

(d)

(e)

(f )

x1 2

(g)

2
x +2

2
+ 1x dx = ln | x 2| + 2( x 2)1
( x 2)2
5
5
dx = 2 ln | x + 2| + x+
2 + c.
( x +2)2

+ ln | x | + c

2. All +c".
(a) 2 arctan( x/5)

(b) 5 ln |25 + x |



x + 5


(c) ln
x 5

You might also like