PARTIAL FRACTIONS AND POLYNOMIAL LONG DIVISION

NOAH WHITE

The basic aim of this note is to describe how to break rational functions into pieces.
For example

x2 − 2x + 3 1 3
=1+ − .
x2 − 1 x−1 x+1

The point is that we don’t know how to integrate the left hand side, but integrating
the right hand side is easy! We will break this problem down into pieces.
In general we have two polynomials p(x) and q(x). Lets say p(x) has degree m
(this means the largest power of x is xm ) and q(x) has degree n. So that

p(x) = pm xm + pm−1 xm−1 + . . . + p1 x + p0

and

q(x) = qn xn + qn−1 xn−1 + . . . + q1 x + q0

for some numbers pi and qi . We want to be able to write

p(x) A1 A2 An
= d(x) + + + ... +
q(x) (a1 x + b1 )ci (a2 x + b2 )c2 (an x + bn )cn

for some numbers Ai and ci and some polynomial r(x). The numbers ai and bi
come from factorising q(x):

q(x) = (a1 x + b1 )(a2 x + b2 ) · · · (an x + bn ).

How are we going to find b(x) and the Ai ’s?

Remark 1. Factorising q(x) into linear factors, as above, might not always be
possible (though it is if we use complex numbers). There is a way to deal with
this however we will sweep this under the rug for now and assume that q(x) can be
neatly factored into linear factors.
1
2 NOAH WHITE

If the above general explanation doesn’t quite make sense, here are some exam-
ples of how we would like to rewrite rational functions:
x−3 7 2
= − (1)
x2 + 3x − 4 5(x + 4) 5(x − 1)

6 1 3 2
= − + (2)
x3 − 8x2 + 19x − 12 x−1 x−3 x−4

x2 + x + 1 1 3 3
= + + (3)
x3 − 3x2 + 3x − 1 x − 1 (x − 1)2 (x − 1)
3

x4 − 3x3 + 12x − 9 103 27

= x2 + 4x + 16 + − (4)
x2 − 7x + 12 x−4 x−3

x4 + 4x2 + x − 5 311 64 1
=x−2+ + + (5)
3 2
x + 2x − 7x + 4 25 (x + 4) 25 (x − 1) 5 (x − 1)2
Hopefully, you will agree that the right hand side of each expression look far easier
to integrate than the left hand side!

1. Polynomial long division

The first step in achieving out aim is to rewrite p(x)/q(x) in the form
p(x) r(x)
= d(x) +
q(x) q(x)
where d(x) and r(x) are polynomials and the degree of r(x) is less than the degree
of q(x). For example, the first step in examples (4) and (5) above would be
x4 − 3x3 + 12x − 9 76x − 201
= x2 + 4x + 16 + 2
x2 − 7x + 12 x − 7x + 12

x4 + 4x2 + x − 5 15x2 − 17x + 3

= x − 2 + .
x3 + 2x2 − 7x + 4 x3 + 2x2 − 7x + 4
This is achieved by polynomial long division. We call d(x) the divisor and r(x) the
remainder.
Polynomial long division works exactly like normal long division:
x2 + 4x + 16
2

4 3
x − 7x + 12 x − 3x + 12x −9
− x4 + 7x3 − 12x2
4x3 − 12x2 + 12x
− 4x3 + 28x2 − 48x
16x2 − 36x − 9
− 16x2 + 112x − 192
76x − 201
 PARTIAL FRACTIONS AND POLYNOMIAL LONG DIVISION 3

Lets annotate this step by step:

x2 − 7x + 12 x4 − 3x3 + 12x −9
First we take the leading term of the numerator x4 − 3x3 + 12x − 9 and divide it
by the leading term of the denominator. So x4 divided by x2 is x2 .
x2


x2 − 7x + 12 x4 − 3x3 + 12x −9
Now we multiply the result, x2 by −1 and the denominator x2 − 7x + 12 and place
it underneath,
x2


x2 − 7x + 12 x4 − 3x3 + 12x −9
− x4 + 7x3 − 12x2

We add the two polynomials together to get

x2
2

4 3
x − 7x + 12 x − 3x + 12x −9
− x4 + 7x3 − 12x2
4x3 − 12x2 + 12x
and repeat the process. Divide the leading term 4x3 by x2 to obtain 4x,
x2 + 4x
2

4 3
x − 7x + 12 x − 3x + 12x −9
− x4 + 7x3 − 12x2
4x3 − 12x2 + 12x
multiply the result by −1 and the denominator and add the resulting polynomial,
x2 + 4x
2

4 3
x − 7x + 12 x − 3x + 12x −9
− x4 + 7x3 − 12x2
4x3 − 12x2 + 12x
− 4x3 + 28x2 − 48x
16x2 − 36x −9
Once more we repeat the process,
x2 + 4x + 16
2

4 3
x − 7x + 12 x − 3x + 12x −9
− x4 + 7x3 − 12x2
4x3 − 12x2 + 12x
− 4x3 + 28x2 − 48x
16x2 − 36x − 9
− 16x2 + 112x − 192
76x − 201
4 NOAH WHITE

Now we are at a point were we cannot repeat the process anymore (since 76x is not
divisible by x2 ) so we halt. The divisor is d(x) = x2 + 4x + 16 and the remainder
is r(x) = 76x − 201.

2. Partial fractions, distinct factors

Now we just need to deal with the case when p(x) has degree less than the degree
of q(x) as in examples (1), (2) and (3). First we will deal with the case where

q(x) = (a1 x + b1 )(a2 x + b2 ) · · · (an x + bn )

and the factors are all distinct. That means we allow q(x) = (x − 1)(x − 2)(x − 3)
but not q(x) = (x − 1)2 (x − 3). In this case it is always true that we can find
constants A1 , A2 , . . . , An such that
p(x) A1 A2 An
= + + ... + .
q(x) a1 x + b1 a2 x + b2 an x + bn
To find these constants we simply multiply out and compare coefficients. We illus-
trate this with an example.

Example 2.
x−3 x−3 A B
= = +
x2 + 3x − 4 (x + 4)(x − 1) x+4 x−1
If we multiply this equation on both sides by (x + 4)(x − 1) we obtain

x − 3 = A(x − 1) + B(x + 4) = (A + B)x − A + 4B.

Since we are comparing two polynomials, the coefficients of every power of x must
be equal. Explicitly, looking at the coefficient of x gives A + B = 1 and looking at
the constant term gives 4B − A = −3. These are simultaneous equations which we
can solve to get
7 2
A= and B = − .
5 5
Example 3.
6 6 A B C
= = + +
x3 − 8x2 + 19x − 12 (x − 1)(x − 3)(x − 4) x−1 x−3 x−4

If we multiply this equation on both sides by (x − 1)(x − 3)(x − 4) we obtain

6 = A(x − 3)(x − 4) + B(x − 1)(x − 4) + C(x − 1)(x − 3)

= (A + B + C)x2 − (7A + 5B + 4C)x + 12A + 4B + 3C.

Looking at the coefficient of x2 gives A + B + C = 0, at the coefficient of x gives

7A + 5B + 4C = 0 and looking at the constant term gives 12A + 4B + 3C = 6.
These are simultaneous equations which we can solve to get

A = 1, B = −3 and C = 2.
 PARTIAL FRACTIONS AND POLYNOMIAL LONG DIVISION 5

3. Partial fraction, distinct quadratic factors

Up to now we have covered only denominators which feature unique linear fac-
tors. When factorising a polynomial, it is also possible that we may obtain quadratic
factors that we cannot factorise further, for example
x3 − x2 + x − 1 = (x2 + 1)(x − 1).
Here x2 + 1 cannot be broken up further into linear factors. So how do we deal
with a fraction of the form
1
?
(x2 + 1)(x − 1)
Answer: for each quadratic factor ax2 + bx + c in the denominator, we get a
summand in the partial fraction expansion of the form
Ax + B
.
ax2 + bx + c
Example 4.
1 Ax + B C
= 2 +
(x2 + 1)(x − 1) x +1 x−1
We multiply both sides by (x2 + 1)(x − 1) to obtain
1 = (Ax + B)(x − 1) + C(x2 + 1).
We can substitute x = 1 to obtain 1 = 2C, i.e. C = 21 . Putting this back into the
equation above and expanding, we get
1 = (Ax + B)(x − 1) + f rac12(x2 + 1)
1 1
= (A + )x2 + (B − A)x − B + .
2 2
1
Equating coefficients we obtain A = B = − 2 . So
1 x+1 1
= + .
(x2 + 1)(x − 1) 2(x2 + 1) 2(x − 1)

4. Partial fractions, repeated factors

We are now left to deal with the case when q(x) has repeated factors. For
example when q(x) = (x − 1)2 (x − 3) or q = (x − 1)3 . In general, for every factor of
q(x) of the form (ax + b)k or (ax2 + bx + c)k , the partial fraction expansion contains
terms of the form
A1 A2 A3 Ak
+ 2
+ 3
+ ··· + .
ax + b (ax + b) (ax + b) (ax + b)k
or
A1 x + B 1 A2 x + B2 A3 x + B3 Ak x + B k
+ + + ··· + .
ax2 + bx + c (ax2 + bx + c)2 (ax2 + bx + c)3 (ax2 + bx + c)k
To find the constants Ai we follow exactly the same process as above. Some exam-
ples should make this clear.
Example 5.
x2 + x + 1 x2 + x + 1 A B C
= = + +
x3 − 3x2 + 3x − 1 (x − 1)3 x − 1 (x − 1)2 (x − 1)3
6 NOAH WHITE

We multiply both sides by (x − 1)3 to obtain

x2 + x + 1 = A(x − 1)2 + B(x − 1) + C
= Ax2 + (−2A + B)x + A − B + C.
Comparing coefficient gives us the simultaneous equations
A=1
−2A + B = 1
A − B + C = 1.
Solving these gives
A = 1, B=3 and C = 3.
Example 6.
15x2 − 17x + 3 15x2 − 17x + 3 A B C
3 2
= 2
= + +
x + 2x − 7x + 4 (x + 4)(x − 1) x + 4 x − 1 (x − 1)2
Multiplying both sides by (x + 4)(x − 1)2 gives
15x2 − 17x + 3 = A(x − 1)2 + B(x − 1)(x + 4) + C(x + 4)
= (A + B)x2 + (−2A + 3B + C)x + A + 4B + 4C.
Comparing coefficient gives us the simultaneous equations
A + B = 15
−2A + 3B + C = −17
A + 4B + 4C = 3.
Solving these gives
311 64 1
A= , B= and C = .
25 25 5
5. Some useful integrals
Here we list some integrals that are useful when using partial fractions to solve
integration questions.
Z
1
dx = ln |x + a| + C
x+a
Z
1 1 1
dx = − +C if n 6= 1
(x + a)n n − 1 (x + a)n−1
Z
1 1 x
2 2
dx = tan−1 +C
a +x a a