1

Lecture 15: Power Series, Taylor Series

In one Pof the previous lectures (Lecture 12), we asked a question whether, for a given x ∈ R,
the series ∞ 1 n
n=0 n! x converges to eP
x . This question will be answered at the end of this lecture.

Observe that the (n + 1)-th term of ∞ 1 n 1 n 1

n=0 n! x is n! x where n! is fixed and x varies in R. Such a
type of series is called a power series. Let us formally define a power series.

Power series

Let an ∈ R for n = 0, 1, 2, ... .PThe series ∞ n

P
n=0 an x , x ∈ R, is called a power series. More
∞ n
generally, if c ∈ R, then the series n=0 an (x − c)P, x ∈ R, is called a power series at c. If we take
y = x − c then the power series at c reduces to ∞ n
n=0 an y which is a power series at 0. In this
lecture we discuss the convergence of power series.

Examples 15.1. 1. Consider the power series ∞ n 1

P
n=0 an x where an = n! for all n. Let us apply
the ratio test and find the set of points in R on which the series converges. For any x ∈ R \ {0},
|an+1 xn+1 | |x|
→ 0 as n → ∞. Therefore, the series ∞ 1 n
P
|an xn | = n+1 n=0 n! x converges absolutely for all
x ∈ R.

2. We know that the geometric series ∞ n

P
n=0 x converges only in (−1, 1).

3. Using the ratio test, it is easy to verify that the series ∞ n

P
n=0 n!x converges only at x = 0.

The following result gives an idea about the set on which a power series converges.

Theorem 15.1. Suppose ∞ n

P
n=0 an x converges for some x0 6= 0 and diverges for some x1 . Then

(i) ∞ n
P
n=0 an x converges absolutely for all x such that |x| < |x0 |;

(ii) ∞ n
P
n=0 an x diverges for all x such that |x| > |x1 |.

Proof (*). (i) Suppose ∞ n n

P
n=0 an x0 converges for some x0 6= 0 and |x| <| x0 |. Since an x0 → 0,
n
there exists M ∈ R such that |an x0 | ≤ M for all n ∈ N. Therefore,
n n
x x
|an xn | = |an xn0 | ≤ M for all n ∈ N.
x0 x0
Since | xx0 | < 1, by the comparison test, the series ∞ n
P
n=0 an x converges absolutely.

(ii) Let x ∈ R and |x| > |x1 |. Suppose ∞

P n
P∞ n
n=0 an x converges. Then by (i), the series n=0 an x1
converges absolutely which is a contradiction. 

For a given power series ∞

P n
P∞ n
n=0 an x , define S = {x ∈ R : n=0 an x is convergent}. If S is
bounded then observe from Theorem 15.1 that sup S can be either 0 or r for some r > 0. Hence,
it follows again from Theorem 15.1 (see Problem 1 of PP15) that the possibilities for S are
{0}, R, (−r, r), [−r, r), (−r, r] and [−r, r] for some r > 0.
If S is R (resp., {0}) then we say that the radius of convergence of the power series is ∞ (resp., 0).
In case S is an interval of the form (−r, r), [−r, r), (−r, r], [−r, r] for some r > 0, then the radius of
convergence is r.

Examples 15.2. 1. We have already seen above that the power series ∞ 1 n
P
n=0 n! x converges
P∞ for all
x ∈ R and hence the radius of convergence is ∞. Similarly, the radius of convergence of n
P∞ n n=0 n!x
(resp., n=0 x ) is 0 (resp., 1).
Please write to [email protected] if any typos/mistakes are found in these notes.
 2

xn
2. Consider the power series ∞
P
n=0 n . Let us apply the ratio test to find the radius of convergence.
Since an = n1 for all n, for x ∈ R \ {0}, we have

|an+1 xn+1 | xn+1 n n

= = x → |x| as n → ∞.
|an xn | (n + 1)xn n+1

It is clear that the series converges absolutely for |x| < 1 and diverges for |x| > 1. Therefore, the
radius of convergence is 1. Since the series converges for x = −1 by the Leibniz test and diverges
for
P∞x =xn1, the set S in this case is [−1, 1). Similarly, the set S corresponding to the power series
n=0 n2 is [−1, 1].

(x−2)n
3. Consider the power series ∞
P
n=0 n3n . We will use the root test to determine the set in which
n
the given series converges. For a given x, let dn = (x−2)
n3n . Then |dn |
1/n → | x−2 |. By the root test,
3
the given series converges for all x satisfying |x − 2| < 3, i.e., −1 < x < 5 and diverges for all x
satisfying |x − 2| > 3. We need to check the convergence/divergence for x = −1 and x = 5. At
x = −1, the series converges by the Leibniz test and the series diverges when x = 5. Hence the
series converges only on [−1, 5). Since the given power series is at 2, the radius of convergence is
considered to be 3.

Let us come back to the question which was posed in the beginning of this lecture.

Taylor’s Series

Taylor’s polynomial which was introduced when we discussed Taylor’s theorem, leads to the
Taylor series. Let D ⊂ R, c ∈ D, f : D → R and f (n) (c) exist for all n. The nth degree polynomial
Pn (x) defined by

f 00 (c) f (n) (c)

Pn (x) = f (c) + f 0 (c)(x − c) + (x − c)2 + · · · + (x − c)n
2! n!
is called the Taylor polynomial (with respect to f and c). The power series
∞
0 f 00 (c) f (n) (c) X f (n) (c)
f (c) + f (c)(x − c) + (x − c)2 + · · · + (x − c)n + · · · (denoted by (x − c)n )
2! n! n!
n=0

is called the Taylor series of f at c. If c = 0, then the Taylor series of f at c is called Maclaurin’s
series.

Remark 15.1. 1. Observe from the definitions of the Taylor polynomial Pn (x) and the Taylor
series (defined above) that Pn (x) is a partial sum of the Taylor series.

2. In order to compute the Taylor series of f at c, we need only f (n) (c) for every n.

ExampleP∞15.3. 1. Let f (x) = ex for all x ∈ R. Since f (n) (x) = ex for all n, the Taylor series of f
1 n
x whereas the Taylor series of f at 1 is ∞ e n
P
at 0 is n=0 n! n=0 n! (x − 1) .
n
2. Let f (x) = 1
6= 0. Observe that for x 6= 0, f (n) (x) = (−1)
x, x xn+1
n!
for all n. It follows that the
P∞ (−1)n ∞ (−1)n
Taylor series of f at 2 is n=0 2n+1 (x − 2) and at 3 is n=0 3n+1 (x − 3)n
n
P

Convergence of Taylor’s series

Taylor’s theorem is needed to discuss the convergence of Taylor’s series. In the rest of this
lecture, we assume f : (a, b) → R to be infinite times differentiable, where a ∈ R ∪ {−∞}, b ∈
 3

R ∪ {∞} and a < b. For every x, c ∈ (a, b) and n ∈ N, by Taylor’s theorem, there exists cn between
x and c such that
f 00 (c) f (n) (c) f (n+1) (cn )
f (x) = f (c) + f 0 (c)(x − c) + (x − c)2 + · · · + (x − c)n + (x − c)n+1 .
2! n! (n + 1)!

One may get an impression from the preceding equation that the Taylor series of f at c, which is
a power series, may converge to f (x) for every x ∈ (a, b). We will see below that the Taylor series
of f may not converge for all x ∈ (a, b) and even if it converges for some x, it need not converge to
f (x).

Example 15.4. 1. If we consider the function P g1 : R\{1} → R given by g1 (x) = 1/(1 − x), then
the Maclaurin series of f is the geometric series ∞ n
n=0 x which converges only on (−1, 1). Observe
that the function g1 is infinite times differentiable on (−∞, 1) but the Maclaurin series of f does
not converge to g1 (x) for any x ∈ (−∞, −1].
2
2. Define g2 (x) = e−1/x for x 6= 0 and g2 (0) = 0. Using the L’Hospital rule, we can show (see
(n)
Problem 5 of PP15) that g2 (0) = 0 for all n = 1, 2, ... Therefore the Maclaurin series of g2 (for
any x ∈ R) is identically zero and it does not converge to g2 (x) at any x 6= 0.

We now answer the important question that under what condition on f , the Taylor series of
f converges to f (x) when x is in the domain of the convergence of the given Taylor’s series. Let
x, c ∈ (a, b). By Taylor’s theorem, there exists cn between x and c such that

f (n+1) (cn )
f (x) − Pn (x) = (x − c)n+1
(n + 1)!
Note that cn depends also on n. Denote En (x) = f (x) − Pn (x). It is clear that the Taylor series
of f at c converges to f (x) if and only if En (x) → 0 (as (Pn (x)) is the sequence of partial sums of
the Taylor series). We use this characterization below to show the convergence of Taylor’s series of
some common functions.

Example 15.5. 1. Let f (x) = sin x, x ∈ R. We will show that the Maclaurin series of f converges
to f (x) at every x ∈ R. Fix x ∈ R. By Taylor’s theorem, there exists cn between 0 and x such that

f (n+1) (cn )
En (x) = (x − 0)n+1 .
(n + 1)!
xn+1 xn
Since | f (n) (t) | ≤ 1 for all n ∈ N and all t ∈ R, observe that |En (x)| ≤ (n+1)! . Let dn = n! for
xn+1
every n. Then | dn+1
dn | → 0. By the ratio test for sequence, | (n+1)! | → 0 and hence |En (x)| → 0.
Therefore, the Macluarin series of f converges to f (x) at all x ∈ R. So, we can expand the sin
(−1)n 2n+1
function in a series form on the whole of R and we write sin x = ∞
P
n=0 (2n+1)! x for every x ∈ R.
P∞ (−1)n 2n
Similarly, we can show that cos x = n=0 (2n)! x for every x ∈ R.

2. We will show that ex = ∞ 1 n x

P
n=0 n! x for all x ∈ R. Let f (x) = e for all x ∈ R. Fix x ∈ R. If
x = 0, then there is nothing to show. Suppose x 6= 0. By Taylor’s Theorem, there exists cn between
0 and x such that
f n+1 (cn ) n+1 ecn
|En (x)| = x = xn+1 .
(n + 1)! (n + 1)!
Since cn depends on n, we bypass cn as follows. If x > 0, then ecn ≤ ex and if x < 0 then ecn ≤ e0 .
e|x| |x|
Therefore, ecn ≤ e|x| and hence |En (x)| ≤ | (n+1)! |xn+1 | for all n. Let dn = en! xn for all n. Then
 4

P∞
| dn+1 x x
dn | = | n+1 | → 0. This implies that dn → 0 and hence En (x) → 0. Therefore, e =
1 n
n=0 n! x
for all x ∈ R.

Convergence of Maclaurin series of ln(1 + x) is discussed in Problems 3 and 6 of PP15.