Taylor Series and Maclaurin Series

Definition (Taylor Series)

Suppose the function f is infinitely di↵erentiable at a.
The Taylor series of f about a (or at a or centered at a) is the power series
1
X f (n) (a) n 0 f 00 (a) 2 f (3) (a) 3
(x a) = f (a) + f (a)(x a) + (x a) + (x a) + ...
n=0
n! 2! 3!
The Taylor series about 0 is called the Maclaurin series.

The definition does not address the convergence issue.

The Taylor series about a is defined as long as derivatives of f of all orders exist at a.
So, what’s so special about the Taylor series? What is its relation with the function f ?

Theorem (Power series representations of functions)

If f has a power series representation about a with a positive (or infinite) radius of convergence R, then that
power series must be the Taylor series of f about a.

Thus, the Taylor series is the only possible candidate for a power series representation of a function.
If it turns out that the Taylor series does not represent (converge to) the function (Note: Yes, this can
happen), no power series can.

Math 267 (University of Calgary) Fall 2015, Winter 2016 1/9

Taylor series representation of functions
Proof of the Theorem: Suppose
1
X n 2 3
f (x) = an (x a) = a0 + a1 (x a) + a2 (x a) + a3 (x a) + ... (c R, c + R)
n=0

f (n) (a)
We want to prove that the coefficients are precisely an = for all n, or, f (n) (a) = n!an .
n!
Letting x = a, we have f (a) = a0 = 0!a0 . (Recall: 0! = 1 by definition, f (0) = f by definition)
By term-by-term di↵erentiation,
0 2 3
f (x) = a1 + 2a2 (x a) + 3a3 (x a) + 4a4 (x a) + ...
Letting x = a, we get f 0 (a) = a1 = 1!a1 .
Di↵erentiate again:
00 2 3
f (x) = 2a2 + 2 · 3a3 (x a) + 3 · 4a4 (x a) + 4 · 5a5 (x a) + ...
Letting x = a, we get f 00 (a) = 2a2 = 2!a2 .
Di↵erentiate again:
(3) 2
f (x) = 2 · 3a3 + 2 · 3 · 4a4 (x a) + 3 · 4 · 5(x a) + ...
Letting x = a, we get f (3) (a) = 2 · 3a3 = 3!a3 .
Continue with this, we get f (n) (a) = 2 · 3 · 4 · ... · nan = n!an .

Math 267 (University of Calgary) Fall 2015, Winter 2016 2/9

Example
Find the Maclaurin series of f (x) = e x . What is its interval of convergence?

Solution: In a previous example, we have proved that

1
X
x xk x2 x3
e = =1+x + + + ... ( 1 < x < 1)
k=0
k! 2! 3!

Since this power series represents e x with radius of convergence R = 1, it must be the Taylor series
(about 0).
The interval of convergence is ( 1, 1).
Alternatively, we can compute the Maclaurin series from definition.
f (n) (x) = e x for all n f (n) (0) = 1 for all n.
So, the Maclaurin series of f (x) = e x is
1
X 1
X
f (n) (0) n xn x2 x3
x = =1+x + + + ...
n=0
n! n=0
n! 2! 3!

This computes the Maclaurin series but does not tell us whether the series converges to e x or not.

Math 267 (University of Calgary) Fall 2015, Winter 2016 3/9

Example
Find the Maclaurin series of f (x) = sin x.

Solution: We make a table:

n f(n) (x) f(n) (0)

0 sin x 0
1 cos x 1
2 -sin x 0
3 -cos x -1
4 sin x 0
5 cos x 1
6 -sin x 0
7 -cos x -1
8 sin x 0
9 cos x 1
. . .
. . .
. . .

The Maclaurin series is

1
X 1
X
f (n) (0) n 1 3 1 5 1 7 ( 1)k 2k+1
x =0+1·x +0 x +0+ x +0 x + ... = x .
n=0
n! 3! 5! 7! k=0
(2k + 1)!

Math 267 (University of Calgary) Fall 2015, Winter 2016 4/9

Taylor Polynomials, Taylor’s Theorem
To study the convergence of the Taylor series, we consider its partial sums:

Definition (Taylor polynomials)

The n-th order Taylor polynomial of f at a is the partial sum of the Taylor series up to the (x a)n term:
n
X f k (a) k 0 f 00 (a) 2 f (n) (a) n
Tn (x) = (x a) = f (a) + f (a)(x a) + (x a) + ... + (x a)
k=0
k! 2! n!

Some authors call Tn (x) the n-th degree Taylor polynomial.

However, the degree of Tn (x) need not be n, since f (n) (a) could be equal to 0.

Theorem (When is f (x) representable as a power series? by its Taylor series?)

Let Tn (x) be the n-th order Taylor polynomial of f at a, and let Rn (x) = f (x) Tn (x). Then the following are
equivalent.
1 f can be represented by a power series about a on the open interval (a R, a + R).
2 The Taylor series of f at a converges to f (x) for all x in (a R, a + R).
3 lim Rn (x) = 0 for all x in (a R, a + R).
n!1

Proof: We have proved (1) =) (2). The rest are straightforward.

Math 267 (University of Calgary) Fall 2015, Winter 2016 5/9

Taylor’s Theorem, Lagrange’s form of the remainder
So, the convergence issue can be resolved by analyzing the remainder term Rn (x).

Theorem (Taylor’s Theorem)

Suppose that f is n + 1 times di↵erentiable on an open interval I containing a. Then for any x in I there is a
number c strictly between a and x such that
f n+1 (c) n+1
Rn (x) = (x a)
(n + 1)!

The above expression for Rn (x) is called the Lagrange’s form of the remainder term.
Taylor’s Theorem can be proved by a method similar to the proof of the Mean Value Theorem. We will
not go into the details.

Corollary (One useful convergence condition)

Suppose that f is infinitely di↵erentiable on an open interval I containing a. If there is a number M such that
k
f (t)  M for all t in I and for all k,

then the Taylor series of f at a converges to f on I .

|x a|n+1 xk
Proof: |Rn (x)|  ! 0 as n ! 1. (Recall a fact: lim = 0 for all x.)
(n + 1)! k!1 k!

Math 267 (University of Calgary) Fall 2015, Winter 2016 6/9

Example
Prove that e x is represented by its Maclaurin series on the interval ( 1, 1).

Solution: Let f (x) = e x .

Take any open interval of the form I = ( A, A), where A > 0.
Then for all t in I and for all k, |f (k) (t)| = |e t | = e t < e A .
By our Corollary, the Maclaurin series of e x converges to e x on the interval ( A, A).
Since A > 0 is arbitrary, the Maclaurin series of e x converges to e x at all points x.

1
X
x xk x2 x3
e = =1+x + + + ... for all x.
k=0
k! 2! 3!

Example
Prove that the Maclaurin series of sin x converges to sin x for all x.

Let f (x) = sin x and a = 0.

f (k) (t) is equal to one of ± sin t, ± cos t.
Since ± sin t and ± cos t is between 1 and 1, f (n) (t)  1.
By our Corollary, the Maclaurin series of sin x converges to sin x on ( 1, 1).

1
X ( 1)k 2k+1 1 3 1 5 1 7
sin x = x =x x + x x + ... for all x.
k=0
(2k + 1)! 3! 5! 7!

Math 267 (University of Calgary) Fall 2015, Winter 2016 7/9

Some important Maclaurin series
Example
Function Maclaurin series representation Radius of Conv.
1
X
1 k k 2 3 4
= ( 1) x = 1 x + x x +x ... 1
1+x k=0

1
X
x xk x2 x3
e = =1+x + + + ... 1
k=0
k! 2! 3!

1
X ( 1)k 2k+1 1 3 1 5 1 7
sin x = x =x x + x x + ... 1
k=0
(2k + 1)! 3! 5! 7!

1
X ( 1)k 2k 1 2 1 4 1 6
cos x = x =1 x + x x + ... 1
k=0
(2k)! 2! 4! 6!

1
X ( 1)k+1 k 1 2 1 3
ln(1 + x) = x =x x + x ... 1
k=1
k 2 3

1
X
1 ( 1)k 2k+1 1 3 1 5 1 7
tan x = x =x x + x x + ... 1
k=0
2k + 1 3 5 7

1 ⇣ ⌘
X
↵ ↵ k ↵(↵ 1) 2 ↵(↵ 1)(↵ 2) 3
(1+x) = x = 1 + ↵x + x + x + ... 1
k=0
k 2! 3!

Math 267 (University of Calgary) Fall 2015, Winter 2016 8/9

Further examples/exercises

Example
1 Verify the Maclaurin series representations on the last slide. Can you verify some of the radii of
convergence?
⇡
2 Find the Taylor series of f (x) = cos x at .
3

Math 267 (University of Calgary) Fall 2015, Winter 2016 9/9