Skip to main content

Maclaurin and Taylor Series

Maclaurin and Taylor Series

Power series provide a way to represent functions as infinite sums, enabling computation, approximation, and analysis that would otherwise be impossible. The Maclaurin series expands a function about x=0x = 0; the Taylor series generalises this to expansion about any point. Together they are among the most powerful tools in analysis and applied mathematics.

Taylor Series Approximation of e^x

Adjust the parameters in the graph above to explore the relationships between variables.

Board Coverage

BoardPaperNotes
AQAPaper 1Maclaurin series; standard series for exe^x, sinx\sin x, cosx\cos x, ln(1+x)\ln(1+x), (1+x)n(1+x)^n
EdexcelFP2Maclaurin + Taylor series about general x=ax = a; convergence
OCR (A)Paper 1Maclaurin series; standard series; applications
CIEP2Maclaurin series; expansion of compound functions; range of validity
All boards require the standard Maclaurin series. Edexcel and CIE additionally require

Taylor series about x=ax = a. The formula booklet lists the standard Maclaurin series — you must be able to derive them and apply them. :::


1. Maclaurin Series

1.1 Definition

Definition. The Maclaurin series of a function ff that is infinitely differentiable at x=0x = 0 is

f(x)=f(0)+xf(0)+x22!f(0)+x33!f(0)+=n=0f(n)(0)n!xn\boxed{f(x) = f(0) + xf'(0) + \frac{x^2}{2!}f''(0) + \frac{x^3}{3!}f'''(0) + \cdots = \sum_{n=0}^{\infty}\frac{f^{(n)}(0)}{n!}\,x^n}

This is a special case of the Taylor series about x=a=0x = a = 0.

1.2 Derivation from the Taylor series

Assume ff can be written as a power series f(x)=a0+a1x+a2x2+a3x3+f(x) = a_0 + a_1 x + a_2 x^2 + a_3 x^3 + \cdots.

Differentiating repeatedly and evaluating at x=0x = 0:

  • f(0)=a0f(0) = a_0
  • f(0)=a1f'(0) = a_1
  • f(0)=2a2f''(0) = 2a_2
  • f(0)=6a3=3!a3f'''(0) = 6a_3 = 3!\,a_3

In general, f(n)(0)=n!anf^{(n)}(0) = n!\,a_n, giving an=f(n)(0)n!a_n = \dfrac{f^{(n)}(0)}{n!}.

Substituting back gives the Maclaurin series.


2. Standard Maclaurin Series

2.1 Exponential function

Proof of the Maclaurin series for exe^x

Let f(x)=exf(x) = e^x. Then f(n)(x)=exf^{(n)}(x) = e^x for all nn, so f(n)(0)=1f^{(n)}(0) = 1 for all nn.

ex=1+x+x22!+x33!+=n=0xnn!e^x = 1 + x + \frac{x^2}{2!} + \frac{x^3}{3!} + \cdots = \sum_{n=0}^{\infty}\frac{x^n}{n!}

ex=n=0xnn!forallxR\boxed{e^x = \sum_{n=0}^{\infty}\frac{x^n}{n!}} \quad \mathrm{for all } x \in \mathbb{R}

\square

Radius of convergence: \infty (converges for all real xx).

2.2 Sine function

Proof of the Maclaurin series for sinx\sin x

Let f(x)=sinxf(x) = \sin x. The derivatives cycle: sinx\sin x, cosx\cos x, sinx-\sin x, cosx-\cos x, sinx\sin x, ...

At x=0x = 0: 0,1,0,1,0,1,0,1,0, 1, 0, -1, 0, 1, 0, -1, \ldots

sinx=xx33!+x55!x77!+=n=0(1)nx2n+1(2n+1)!\sin x = x - \frac{x^3}{3!} + \frac{x^5}{5!} - \frac{x^7}{7!} + \cdots = \sum_{n=0}^{\infty}\frac{(-1)^n x^{2n+1}}{(2n+1)!}

sinx=n=0(1)nx2n+1(2n+1)!forallxR\boxed{\sin x = \sum_{n=0}^{\infty}\frac{(-1)^n x^{2n+1}}{(2n+1)!}} \quad \mathrm{for all } x \in \mathbb{R}

\square

Radius of convergence: \infty.

2.3 Cosine function

Let f(x)=cosxf(x) = \cos x. The derivatives cycle: cosx\cos x, sinx-\sin x, cosx-\cos x, sinx\sin x, cosx\cos x, ...

At x=0x = 0: 1,0,1,0,1,0,1,0,1, 0, -1, 0, 1, 0, -1, 0, \ldots

cosx=1x22!+x44!x66!+=n=0(1)nx2n(2n)!forallxR\boxed{\cos x = 1 - \frac{x^2}{2!} + \frac{x^4}{4!} - \frac{x^6}{6!} + \cdots = \sum_{n=0}^{\infty}\frac{(-1)^n x^{2n}}{(2n)!}} \quad \mathrm{for all } x \in \mathbb{R}

2.4 Natural logarithm

Let f(x)=ln(1+x)f(x) = \ln(1+x) for x<1|x| < 1.

f(x)=11+xf'(x) = \dfrac{1}{1+x}, f(x)=1(1+x)2f''(x) = -\dfrac{1}{(1+x)^2}, f(x)=2(1+x)3f'''(x) = \dfrac{2}{(1+x)^3}, ...

f(n)(x)=(1)n1(n1)!(1+x)nf^{(n)}(x) = \dfrac{(-1)^{n-1}(n-1)!}{(1+x)^n}, so f(n)(0)=(1)n1(n1)!f^{(n)}(0) = (-1)^{n-1}(n-1)!.

ln(1+x)=xx22+x33x44+=n=1(1)n1xnnfor1<x1\boxed{\ln(1+x) = x - \frac{x^2}{2} + \frac{x^3}{3} - \frac{x^4}{4} + \cdots = \sum_{n=1}^{\infty}\frac{(-1)^{n-1}x^n}{n}} \quad \mathrm{for } -1 < x \leq 1

Radius of convergence: 1. The series also converges at x=1x = 1 (alternating harmonic series).

2.5 Binomial series

For x<1|x| < 1 and any real nn:

(1+x)n=1+nx+n(n1)2!x2+n(n1)(n2)3!x3+=k=0(nk)xk\boxed{(1+x)^n = 1 + nx + \frac{n(n-1)}{2!}x^2 + \frac{n(n-1)(n-2)}{3!}x^3 + \cdots = \sum_{k=0}^{\infty}\binom{n}{k}x^k}

where (nk)=LBn(n1)(n2)(nk+1)RB◆◆LBk!RB\dbinom{n}{k} = \dfrac◆LB◆n(n-1)(n-2)\cdots(n-k+1)◆RB◆◆LB◆k!◆RB◆.

Radius of convergence: 1.

2.6 Arctangent

arctanx=xx33+x55x77+=n=0(1)nx2n+12n+1forx1\boxed{\arctan x = x - \frac{x^3}{3} + \frac{x^5}{5} - \frac{x^7}{7} + \cdots = \sum_{n=0}^{\infty}\frac{(-1)^n x^{2n+1}}{2n+1}} \quad \mathrm{for } |x| \leq 1

Radius of convergence: 1. Converges at x=±1x = \pm 1 by the alternating series test.

Derivation. Since ddxarctanx=11+x2=1x2+x4x6+\dfrac{d}{dx}\arctan x = \dfrac{1}{1+x^2} = 1 - x^2 + x^4 - x^6 + \cdots for x<1|x| < 1, integrate term by term.

2.7 Summary table

| Function | Maclaurin Series | Valid for | | ----------- | ----------------------------------------------------------------- | --------------- | --- | ------- | | exe^x | n=0xnn!\displaystyle\sum_{n=0}^{\infty}\frac{x^n}{n!} | all xx | | sinx\sin x | n=0(1)nx2n+1(2n+1)!\displaystyle\sum_{n=0}^{\infty}\frac{(-1)^n x^{2n+1}}{(2n+1)!} | all xx | | cosx\cos x | n=0(1)nx2n(2n)!\displaystyle\sum_{n=0}^{\infty}\frac{(-1)^n x^{2n}}{(2n)!} | all xx | | ln(1+x)\ln(1+x) | n=1(1)n1xnn\displaystyle\sum_{n=1}^{\infty}\frac{(-1)^{n-1}x^n}{n} | 1<x1-1 < x \leq 1 | | (1+x)n(1+x)^n | k=0(nk)xk\displaystyle\sum_{k=0}^{\infty}\binom{n}{k}x^k | x<1 | x | < 1 | | arctanx\arctan x | n=0(1)nx2n+12n+1\displaystyle\sum_{n=0}^{\infty}\frac{(-1)^n x^{2n+1}}{2n+1} | x1 | x | \leq 1 |


3. Taylor Series About x=ax = a

3.1 Definition

Definition. The Taylor series of ff about x=ax = a is

f(x)=f(a)+(xa)f(a)+(xa)22!f(a)+(xa)33!f(a)+=n=0f(n)(a)n!(xa)n\boxed{f(x) = f(a) + (x-a)f'(a) + \frac{(x-a)^2}{2!}f''(a) + \frac{(x-a)^3}{3!}f'''(a) + \cdots = \sum_{n=0}^{\infty}\frac{f^{(n)}(a)}{n!}(x-a)^n}

The Maclaurin series is the special case a=0a = 0.

3.2 Taylor's theorem with remainder

Theorem. If ff is (n+1)(n+1)-times differentiable on an interval containing aa and xx, then

f(x)=k=0nf(k)(a)k!(xa)k+Rnf(x) = \sum_{k=0}^{n}\frac{f^{(k)}(a)}{k!}(x-a)^k + R_n

where the remainder in Lagrange form is

Rn=f(n+1)(c)(n+1)!(xa)n+1\boxed{R_n = \frac{f^{(n+1)}(c)}{(n+1)!}(x-a)^{n+1}}

for some cc between aa and xx.

Intuition. The remainder term tells us the error when truncating the series after nn terms. If Rn0|R_n| \to 0 as nn \to \infty, then the Taylor series converges to f(x)f(x).

3.3 Worked example

Find the Taylor series of exe^x about x=1x = 1 up to the x3x^3 term.

f(x)=exf(x) = e^x, f(x)=exf'(x) = e^x, f(x)=exf''(x) = e^x, f(x)=exf'''(x) = e^x.

At x=1x = 1: f(1)=f(1)=f(1)=f(1)=ef(1) = f'(1) = f''(1) = f'''(1) = e.

ex=e+e(x1)+e2(x1)2+e6(x1)3+=en=0(x1)nn!e^x = e + e(x-1) + \frac{e}{2}(x-1)^2 + \frac{e}{6}(x-1)^3 + \cdots = e\sum_{n=0}^{\infty}\frac{(x-1)^n}{n!}

3.4 Converting between Maclaurin and Taylor series

Example. Find the Taylor series of lnx\ln x about x=1x = 1.

Let u=x1u = x - 1, so lnx=ln(1+u)\ln x = \ln(1 + u).

ln(1+u)=uu22+u33\ln(1+u) = u - \frac{u^2}{2} + \frac{u^3}{3} - \cdots

lnx=(x1)(x1)22+(x1)33for0<x2\boxed{\ln x = (x-1) - \frac{(x-1)^2}{2} + \frac{(x-1)^3}{3} - \cdots} \quad \mathrm{for } 0 < x \leq 2


4. Interval and Radius of Convergence

4.1 Definitions

Definition. The radius of convergence RR of a power series n=0anxn\displaystyle\sum_{n=0}^{\infty}a_n x^n is the value such that the series converges for x<R|x| < R and diverges for x>R|x| > R.

The interval of convergence is the set of all xx for which the series converges. The endpoints x=±Rx = \pm R must be tested separately.

4.2 The ratio test

Theorem (Ratio test). For the series n=0anxn\displaystyle\sum_{n=0}^{\infty}a_n x^n, let

L=limnan+1anL = \lim_{n\to\infty}\left|\frac{a_{n+1}}{a_n}\right|

If LL exists, the radius of convergence is R=1LR = \dfrac{1}{L}.

Equivalently, if L=limnan+1xn+1anxn=xlimnan+1anL = \lim_{n\to\infty}\left|\dfrac{a_{n+1}x^{n+1}}{a_n x^n}\right| = |x|\lim_{n\to\infty}\left|\dfrac{a_{n+1}}{a_n}\right|:

  • If L<1L < 1: the series converges absolutely.
  • If L>1L > 1: the series diverges.
  • If L=1L = 1: the test is inconclusive.

4.3 Worked examples

Example 1. Find the radius of convergence of n=1(1)n1xnn\displaystyle\sum_{n=1}^{\infty}\frac{(-1)^{n-1}x^n}{n}.

an=(1)n1na_n = \dfrac{(-1)^{n-1}}{n}, an+1=(1)nn+1a_{n+1} = \dfrac{(-1)^n}{n+1}.

an+1an=nn+11\left|\dfrac{a_{n+1}}{a_n}\right| = \dfrac{n}{n+1} \to 1 as nn \to \infty.

R=1R = 1. The series converges for x<1|x| < 1.

At x=1x = 1: alternating harmonic series, converges. At x=1x = -1: 1/n-\sum 1/n, diverges.

Interval of convergence: 1<x1-1 < x \leq 1.

Example 2. Find the radius of convergence of n=0xnn!\displaystyle\sum_{n=0}^{\infty}\frac{x^n}{n!}.

an+1an=1/(n+1)!1/n!=1n+10\left|\dfrac{a_{n+1}}{a_n}\right| = \dfrac{1/(n+1)!}{1/n!} = \dfrac{1}{n+1} \to 0.

R=R = \infty. Converges for all xx.

4.4 Convergence of the geometric series

Proof of convergence of the geometric series

For x<1|x| < 1, the geometric series is n=0xn\displaystyle\sum_{n=0}^{\infty}x^n.

The partial sum is SN=1+x+x2++xNS_N = 1 + x + x^2 + \cdots + x^N.

SN=1xN+11xS_N = \dfrac{1 - x^{N+1}}{1 - x} (standard formula).

Since x<1|x| < 1: xN+10x^{N+1} \to 0 as NN \to \infty, so SN11xS_N \to \dfrac{1}{1-x}.

n=0xn=11xforx<1\boxed{\sum_{n=0}^{\infty}x^n = \frac{1}{1-x} \quad \mathrm{for } |x| < 1}

\square

tip 11+x\dfrac{1}{1+x}, ln(1+x)\ln(1+x), and arctanx\arctan x all derive from it. :::

5. Applications

5.1 Approximating values

Example. Approximate e0.1e^{0.1} using the first four terms of the Maclaurin series.

e0.11+0.1+0.012+0.0016=1+0.1+0.005+0.0001671.10517e^{0.1} \approx 1 + 0.1 + \dfrac{0.01}{2} + \dfrac{0.001}{6} = 1 + 0.1 + 0.005 + 0.000167 \approx 1.10517.

The actual value is 1.10517\approx 1.10517, so the error is negligible with just 4 terms.

Example. Approximate sin(0.1)\sin(0.1) and bound the error.

sin(0.1)0.10.0016=0.10.000167=0.099833\sin(0.1) \approx 0.1 - \dfrac{0.001}{6} = 0.1 - 0.000167 = 0.099833.

The next term is (0.1)51208.3×108\dfrac{(0.1)^5}{120} \approx 8.3 \times 10^{-8}, so the error is at most this.

5.2 Evaluating limits using series

Example. Find limx0ex1xx2\displaystyle\lim_{x\to 0}\frac{e^x - 1 - x}{x^2}.

ex=1+x+x22+x36+e^x = 1 + x + \dfrac{x^2}{2} + \dfrac{x^3}{6} + \cdots

ex1x=x22+x36+e^x - 1 - x = \dfrac{x^2}{2} + \dfrac{x^3}{6} + \cdots

ex1xx2=12+x6+12\dfrac{e^x - 1 - x}{x^2} = \dfrac{1}{2} + \dfrac{x}{6} + \cdots \to \dfrac{1}{2} as x0x \to 0.

5.3 Series expansions of compound functions

Example. Find the Maclaurin series of ex2e^{x^2} up to the x6x^6 term.

Substitute x2x^2 into the series for exe^x:

ex2=1+x2+x42!+x63!+=1+x2+x42+x66+e^{x^2} = 1 + x^2 + \frac{x^4}{2!} + \frac{x^6}{3!} + \cdots = 1 + x^2 + \frac{x^4}{2} + \frac{x^6}{6} + \cdots

Example. Find the Maclaurin series of cos(x2)\cos(x^2) up to the x8x^8 term.

cos(x2)=1x42!+x84!=1x42+x824\cos(x^2) = 1 - \frac{x^4}{2!} + \frac{x^8}{4!} - \cdots = 1 - \frac{x^4}{2} + \frac{x^8}{24} - \cdots

Example. Find ex1x\dfrac{e^x}{1-x} up to the x3x^3 term.

ex=1+x+x22+x36+e^x = 1 + x + \dfrac{x^2}{2} + \dfrac{x^3}{6} + \cdots

11x=1+x+x2+x3+\dfrac{1}{1-x} = 1 + x + x^2 + x^3 + \cdots

Multiply (Cauchy product):

ex1x=(1+x+x22+x36)(1+x+x2+x3)+O(x4)\dfrac{e^x}{1-x} = \left(1 + x + \dfrac{x^2}{2} + \dfrac{x^3}{6}\right)\left(1 + x + x^2 + x^3\right) + O(x^4)

=1+(1+1)x+(1+1+12)x2+(1+1+12+16)x3= 1 + (1+1)x + \left(1+1+\dfrac{1}{2}\right)x^2 + \left(1+1+\dfrac{1}{2}+\dfrac{1}{6}\right)x^3

=1+2x+52x2+83x3+= 1 + 2x + \dfrac{5}{2}x^2 + \dfrac{8}{3}x^3 + \cdots

5.4 L'Hôpital's rule and series

Intuition. L'Hôpital's rule is a consequence of Taylor's theorem. If f(a)=g(a)=0f(a) = g(a) = 0, then near aa:

f(x)g(x)LBf(a)(xa)+f(a)2(xa)2RB◆◆LBg(a)(xa)+g(a)2(xa)2RB=LBf(a)+f(a)2(xa)RB◆◆LBg(a)+g(a)2(xa)RB\frac{f(x)}{g(x)} \approx \frac◆LB◆f'(a)(x-a) + \frac{f''(a)}{2}(x-a)^2◆RB◆◆LB◆g'(a)(x-a) + \frac{g''(a)}{2}(x-a)^2◆RB◆ = \frac◆LB◆f'(a) + \frac{f''(a)}{2}(x-a)◆RB◆◆LB◆g'(a) + \frac{g''(a)}{2}(x-a)◆RB◆

Taking xax \to a gives f(a)g(a)\dfrac{f'(a)}{g'(a)}, which is L'Hôpital's rule. The series approach often gives more information (higher-order terms) than applying L'Hôpital's rule repeatedly.


6. Key Results and Pitfalls

Common errors:
  1. Wrong coefficients. The Maclaurin coefficient of xnx^n is f(n)(0)n!\dfrac{f^{(n)}(0)}{n!}, not f(n)(0)f^{(n)}(0). Always divide by the factorial.

  2. Forgetting the range of validity. ln(1+x)\ln(1+x) is valid for 1<x1-1 < x \leq 1, not all xx. Using it outside this range gives a wrong answer.

  3. Incorrect substitution. When expanding e2xe^{2x}, substitute 2x2x into every term: e2x=1+2x+4x22+8x36+e^{2x} = 1 + 2x + \dfrac{4x^2}{2} + \dfrac{8x^3}{6} + \cdots, not e2x=1+2x+x22+e^{2x} = 1 + 2x + \dfrac{x^2}{2} + \cdots.

  4. Sign errors in alternating series. sinx\sin x and cosx\cos x have alternating signs: sinx=xx36+\sin x = x - \dfrac{x^3}{6} + \cdots (not x+x36+x + \dfrac{x^3}{6} + \cdots). :::

Exam strategy:
  • Memorise the six standard series. Derive others from them by substitution and algebra.
  • For compound functions (ex2sinxe^{x^2}\sin x), multiply series term by term and collect like powers.
  • Always state the range of validity when asked.
  • To find a Maclaurin series efficiently, compute successive derivatives at x=0x = 0 and look for the pattern. :::

Problems

Problem 1

Find the Maclaurin series of f(x)=cos2xf(x) = \cos 2x up to the x6x^6 term.

Solution 1

f(0)=1f(0) = 1, f(x)=2sin2xf'(x) = -2\sin 2x, f(0)=0f'(0) = 0, f(x)=4cos2xf''(x) = -4\cos 2x, f(0)=4f''(0) = -4, f(x)=8sin2xf'''(x) = 8\sin 2x, f(0)=0f'''(0) = 0, f(4)(x)=16cos2xf^{(4)}(x) = 16\cos 2x, f(4)(0)=16f^{(4)}(0) = 16, f(5)(0)=0f^{(5)}(0) = 0, f(6)(0)=64f^{(6)}(0) = -64.

Alternatively, substitute 2x2x into cosx=1x22!+x44!x66!+\cos x = 1 - \dfrac{x^2}{2!} + \dfrac{x^4}{4!} - \dfrac{x^6}{6!} + \cdots:

cos2x=12x2+2x434x645+\boxed{\cos 2x = 1 - 2x^2 + \frac{2x^4}{3} - \frac{4x^6}{45} + \cdots}

If you get this wrong, revise: Standard Series — Section 2.

Problem 2

Find the Maclaurin series of ln(1x2)\ln(1 - x^2) up to the x6x^6 term, stating the range of validity.

Solution 2

Substitute x2-x^2 into ln(1+u)=uu22+u33\ln(1+u) = u - \dfrac{u^2}{2} + \dfrac{u^3}{3} - \cdots:

ln(1x2)=x2x42x63\ln(1 - x^2) = -x^2 - \frac{x^4}{2} - \frac{x^6}{3} - \cdots

Range: x2<1    x<1|x^2| < 1 \implies |x| < 1. Also converges at x=±1x = \pm 1 (alternating harmonic at x=1x = 1 when substituted, but here x2x4/2x6/3-x^2 - x^4/2 - x^6/3 - \cdots at x=1x = 1 is 1/n-\sum 1/n, diverges).

Valid for 1<x<1-1 < x < 1.

If you get this wrong, revise: Standard Series — Section 2.

Problem 3

Use the Maclaurin series for exe^x to show that n=02nn!=e2\displaystyle\sum_{n=0}^{\infty}\frac{2^n}{n!} = e^2.

Solution 3

ex=n=0xnn!e^x = \displaystyle\sum_{n=0}^{\infty}\dfrac{x^n}{n!}.

Setting x=2x = 2: e2=n=02nn!e^2 = \displaystyle\sum_{n=0}^{\infty}\dfrac{2^n}{n!}. \blacksquare

If you get this wrong, revise: Exponential Function — Section 2.1.

Problem 4

Find the Taylor series of sinx\sin x about x=LBπRB◆◆LB3RBx = \dfrac◆LB◆\pi◆RB◆◆LB◆3◆RB◆ up to the (xπ/3)3(x - \pi/3)^3 term.

Solution 4

f(x)=sinxf(x) = \sin x. f(π/3)=LB3RB◆◆LB2RBf(\pi/3) = \dfrac◆LB◆\sqrt{3}◆RB◆◆LB◆2◆RB◆. f(x)=cosxf'(x) = \cos x, f(π/3)=12f'(\pi/3) = \dfrac{1}{2}. f(x)=sinxf''(x) = -\sin x, f(π/3)=LB3RB◆◆LB2RBf''(\pi/3) = -\dfrac◆LB◆\sqrt{3}◆RB◆◆LB◆2◆RB◆. f(x)=cosxf'''(x) = -\cos x, f(π/3)=12f'''(\pi/3) = -\dfrac{1}{2}.

sinx=LB3RB◆◆LB2RB+12(xLBπRB◆◆LB3RB)LB3RB◆◆LB4RB(xLBπRB◆◆LB3RB)2112(xLBπRB◆◆LB3RB)3+\sin x = \frac◆LB◆\sqrt{3}◆RB◆◆LB◆2◆RB◆ + \frac{1}{2}\left(x-\frac◆LB◆\pi◆RB◆◆LB◆3◆RB◆\right) - \frac◆LB◆\sqrt{3}◆RB◆◆LB◆4◆RB◆\left(x-\frac◆LB◆\pi◆RB◆◆LB◆3◆RB◆\right)^2 - \frac{1}{12}\left(x-\frac◆LB◆\pi◆RB◆◆LB◆3◆RB◆\right)^3 + \cdots

If you get this wrong, revise: Taylor Series About x=ax = a — Section 3.

Problem 5

Use series to evaluate limx0LBsinxxRB◆◆LBx3RB\displaystyle\lim_{x\to 0}\frac◆LB◆\sin x - x◆RB◆◆LB◆x^3◆RB◆.

Solution 5

sinx=xx36+x5120\sin x = x - \dfrac{x^3}{6} + \dfrac{x^5}{120} - \cdots

sinxx=x36+x5120\sin x - x = -\dfrac{x^3}{6} + \dfrac{x^5}{120} - \cdots

LBsinxxRB◆◆LBx3RB=16+x212016\dfrac◆LB◆\sin x - x◆RB◆◆LB◆x^3◆RB◆ = -\dfrac{1}{6} + \dfrac{x^2}{120} - \cdots \to \boxed{-\dfrac{1}{6}} as x0x \to 0.

If you get this wrong, revise: Evaluating Limits — Section 5.2.

Problem 6

Find the Maclaurin series of (1+2x)1/2(1 + 2x)^{-1/2} up to the x3x^3 term and state the range of validity.

Solution 6

Using the binomial series with n=1/2n = -1/2:

(1+u)1/2=1+(12)u+(1/2)(3/2)2u2+(1/2)(3/2)(5/2)6u3+(1+u)^{-1/2} = 1 + \left(-\dfrac{1}{2}\right)u + \dfrac{(-1/2)(-3/2)}{2}u^2 + \dfrac{(-1/2)(-3/2)(-5/2)}{6}u^3 + \cdots

=1u2+3u285u316+= 1 - \dfrac{u}{2} + \dfrac{3u^2}{8} - \dfrac{5u^3}{16} + \cdots

Substituting u=2xu = 2x:

(1+2x)1/2=1x+3x225x32+\boxed{(1+2x)^{-1/2} = 1 - x + \frac{3x^2}{2} - \frac{5x^3}{2} + \cdots}

Valid for 2x<1    x<12|2x| < 1 \implies |x| < \dfrac{1}{2}.

If you get this wrong, revise: Binomial Series — Section 2.5.

Problem 7

Find the radius of convergence of n=0(2x)nn2+1\displaystyle\sum_{n=0}^{\infty}\frac{(2x)^n}{n^2 + 1}.

Solution 7

an=2nn2+1a_n = \dfrac{2^n}{n^2+1}.

an+1an=2n+1(n2+1)2n(n2+2n+2)=2(n2+1)n2+2n+22\left|\dfrac{a_{n+1}}{a_n}\right| = \dfrac{2^{n+1}(n^2+1)}{2^n(n^2+2n+2)} = \dfrac{2(n^2+1)}{n^2+2n+2} \to 2 as nn \to \infty.

R=12R = \dfrac{1}{2}. The series converges for x<12\boxed{|x| < \dfrac{1}{2}}.

If you get this wrong, revise: The Ratio Test — Section 4.2.

Problem 8

Find the Maclaurin series of exsinxe^x \sin x up to the x5x^5 term.

Solution 8

ex=1+x+x22+x36+x424+e^x = 1 + x + \dfrac{x^2}{2} + \dfrac{x^3}{6} + \dfrac{x^4}{24} + \cdots

sinx=xx36+x5120\sin x = x - \dfrac{x^3}{6} + \dfrac{x^5}{120} - \cdots

Multiply and collect terms:

xx: 1x=x1 \cdot x = x

x2x^2: xx=x2x \cdot x = x^2

x3x^3: x22x+1(x36)=x32x36=x33\dfrac{x^2}{2} \cdot x + 1 \cdot \left(-\dfrac{x^3}{6}\right) = \dfrac{x^3}{2} - \dfrac{x^3}{6} = \dfrac{x^3}{3}

x4x^4: x36x+x(x36)=0\dfrac{x^3}{6} \cdot x + x \cdot \left(-\dfrac{x^3}{6}\right) = 0

x5x^5: x424x+x22(x36)+1x5120=x524x512+x5120=510+1120x5=x530\dfrac{x^4}{24} \cdot x + \dfrac{x^2}{2}\left(-\dfrac{x^3}{6}\right) + 1 \cdot \dfrac{x^5}{120} = \dfrac{x^5}{24} - \dfrac{x^5}{12} + \dfrac{x^5}{120} = \dfrac{5 - 10 + 1}{120}x^5 = -\dfrac{x^5}{30}

exsinx=x+x2+x33x530+\boxed{e^x \sin x = x + x^2 + \frac{x^3}{3} - \frac{x^5}{30} + \cdots}

If you get this wrong, revise: Compound Functions — Section 5.3.

Problem 9

Show that n=0(1)n(2n+1)!=sin1\displaystyle\sum_{n=0}^{\infty}\frac{(-1)^n}{(2n+1)!} = \sin 1.

Solution 9

sinx=n=0(1)nx2n+1(2n+1)!\sin x = \displaystyle\sum_{n=0}^{\infty}\dfrac{(-1)^n x^{2n+1}}{(2n+1)!}.

Setting x=1x = 1: sin1=n=0(1)n(2n+1)!\sin 1 = \displaystyle\sum_{n=0}^{\infty}\dfrac{(-1)^n}{(2n+1)!}. \blacksquare

If you get this wrong, revise: Sine Function — Section 2.2.

Problem 10

Use series to evaluate limx0ex21x2\displaystyle\lim_{x\to 0}\frac{e^{-x^2} - 1}{x^2}.

Solution 10

ex2=1x2+x42e^{-x^2} = 1 - x^2 + \dfrac{x^4}{2} - \cdots

ex21=x2+x42e^{-x^2} - 1 = -x^2 + \dfrac{x^4}{2} - \cdots

ex21x2=1+x221\dfrac{e^{-x^2} - 1}{x^2} = -1 + \dfrac{x^2}{2} - \cdots \to \boxed{-1} as x0x \to 0.

If you get this wrong, revise: Evaluating Limits — Section 5.2.


7. Advanced Worked Examples

Example 7.1: Maclaurin series of a composite function

Problem. Find the Maclaurin series of f(x)=LB1RB◆◆LB1+x2RBf(x) = \dfrac◆LB◆1◆RB◆◆LB◆\sqrt{1 + x^2}◆RB◆ up to the x6x^6 term.

Solution. Use the binomial series with n=1/2n = -1/2:

(1+u)1/2=1u2+3u285u316+35u4128(1 + u)^{-1/2} = 1 - \dfrac{u}{2} + \dfrac{3u^2}{8} - \dfrac{5u^3}{16} + \dfrac{35u^4}{128} - \cdots

Substituting u=x2u = x^2:

LB1RB◆◆LB1+x2RB=1x22+3x485x616+\frac◆LB◆1◆RB◆◆LB◆\sqrt{1+x^2}◆RB◆ = 1 - \frac{x^2}{2} + \frac{3x^4}{8} - \frac{5x^6}{16} + \cdots

Valid for x2<1|x^2| < 1, i.e., x<1|x| < 1.

Example 7.2: Taylor series and error bounds

Problem. Use the Taylor series of cosx\cos x about x=0x = 0 to approximate cos(0.2)\cos(0.2) and bound the error.

Solution. cosx=1x22+x424x6720+\cos x = 1 - \dfrac{x^2}{2} + \dfrac{x^4}{24} - \dfrac{x^6}{720} + \cdots

Using three terms: cos(0.2)10.02+0.001624=10.02+0.0000667=0.980067\cos(0.2) \approx 1 - 0.02 + \dfrac{0.0016}{24} = 1 - 0.02 + 0.0000667 = 0.980067.

Error bound: the next term is (0.2)6720=LB6.4×107RB◆◆LB720RB8.9×1010\dfrac{(0.2)^6}{720} = \dfrac◆LB◆6.4 \times 10^{-7}◆RB◆◆LB◆720◆RB◆ \approx 8.9 \times 10^{-10}.

So error<109|\text{error}| < 10^{-9}.

Example 7.3: Product of two Maclaurin series

Problem. Find the Maclaurin series of excosxe^x \cos x up to the x4x^4 term.

Solution. ex=1+x+x22+x36+x424+e^x = 1 + x + \dfrac{x^2}{2} + \dfrac{x^3}{6} + \dfrac{x^4}{24} + \cdots

cosx=1x22+x424\cos x = 1 - \dfrac{x^2}{2} + \dfrac{x^4}{24} - \cdots

Multiplying and collecting terms:

11: 1×1=11 \times 1 = 1 xx: 1x=x1 \cdot x = x x2x^2: 1(x2/2)+xx=x2/2+x2=x2/21 \cdot (-x^2/2) + x \cdot x = -x^2/2 + x^2 = x^2/2 x3x^3: 10+x(x2/2)+(x2/2)x=x3/2+x3/2=01 \cdot 0 + x \cdot (-x^2/2) + (x^2/2) \cdot x = -x^3/2 + x^3/2 = 0 x4x^4: 1(x4/24)+x0+(x2/2)(x2/2)+(x3/6)x=x4/24x4/4+x4/6=(16+4)x4/24=x4/241 \cdot (x^4/24) + x \cdot 0 + (x^2/2)(-x^2/2) + (x^3/6) \cdot x = x^4/24 - x^4/4 + x^4/6 = (1 - 6 + 4)x^4/24 = -x^4/24

excosx=1+x+x22x424+e^x\cos x = 1 + x + \frac{x^2}{2} - \frac{x^4}{24} + \cdots

Example 7.4: Series expansion of an inverse function

Problem. Find the Maclaurin series of secx\sec x up to the x4x^4 term.

Solution. Write secx=LB1RB◆◆LBcosxRB=(1x2/2+x4/24)1\sec x = \dfrac◆LB◆1◆RB◆◆LB◆\cos x◆RB◆ = (1 - x^2/2 + x^4/24 - \cdots)^{-1}.

Using (1u)1=1+u+u2+(1 - u)^{-1} = 1 + u + u^2 + \cdots with u=x2/2x4/24+u = x^2/2 - x^4/24 + \cdots:

secx=1+(x22x424)+(x22)2+\sec x = 1 + \left(\frac{x^2}{2} - \frac{x^4}{24}\right) + \left(\frac{x^2}{2}\right)^2 + \cdots

=1+x22+(x424+x44)+=1+x22+5x424+= 1 + \frac{x^2}{2} + \left(-\frac{x^4}{24} + \frac{x^4}{4}\right) + \cdots = 1 + \frac{x^2}{2} + \frac{5x^4}{24} + \cdots

Example 7.5: Using series to find a limit

Problem. Evaluate limx0LBxsinxRB◆◆LBx3RB\displaystyle\lim_{x\to 0}\frac◆LB◆x - \sin x◆RB◆◆LB◆x^3◆RB◆.

Solution. sinx=xx36+x5120\sin x = x - \dfrac{x^3}{6} + \dfrac{x^5}{120} - \cdots

LBxsinxRB◆◆LBx3RB=LBxx+x3/6x5/120+RB◆◆LBx3RB=16x2120+16\frac◆LB◆x - \sin x◆RB◆◆LB◆x^3◆RB◆ = \frac◆LB◆x - x + x^3/6 - x^5/120 + \cdots◆RB◆◆LB◆x^3◆RB◆ = \frac{1}{6} - \frac{x^2}{120} + \cdots \to \boxed{\frac{1}{6}}

Example 7.6: Maclaurin series by differentiation

Problem. Find the Maclaurin series of f(x)=(1+x)3exf(x) = (1 + x)^3 e^x up to the x3x^3 term.

Solution. f(x)=(1+3x+3x2+x3)exf(x) = (1 + 3x + 3x^2 + x^3)e^x.

ex=1+x+x2/2+x3/6+e^x = 1 + x + x^2/2 + x^3/6 + \cdots

f(x)=(1+3x+3x2+x3)(1+x+x2/2+x3/6)f(x) = (1 + 3x + 3x^2 + x^3)(1 + x + x^2/2 + x^3/6)

11: 1×1=11 \times 1 = 1 xx: 1+3=41 + 3 = 4 x2x^2: 1/2+3+3=13/21/2 + 3 + 3 = 13/2 x3x^3: 1/6+3/2+3+1=1/6+9/6+18/6+6/6=34/6=17/31/6 + 3/2 + 3 + 1 = 1/6 + 9/6 + 18/6 + 6/6 = 34/6 = 17/3

f(x)=1+4x+13x22+17x33+f(x) = 1 + 4x + \frac{13x^2}{2} + \frac{17x^3}{3} + \cdots

Example 7.7: Convergence of alternating series

Problem. How many terms of n=1(1)n+1n\displaystyle\sum_{n=1}^{\infty}\frac{(-1)^{n+1}}{n} are needed to approximate ln2\ln 2 with an error less than 0.0010.001?

Solution. This is the alternating harmonic series, with S=ln2S = \ln 2. The error after NN terms is bounded by the absolute value of the (N+1)(N+1)-th term:

error1N+1<0.001    N+1>1000    N1000|\text{error}| \leq \frac{1}{N+1} < 0.001 \implies N + 1 > 1000 \implies N \geq 1000

So at least 1000 terms are needed.

Example 7.8: Taylor series and approximation of definite integrals

Problem. Use the Maclaurin series of ex2e^{-x^2} to approximate 01/2ex2dx\displaystyle\int_0^{1/2} e^{-x^2}\,dx to 5 decimal places.

Solution. ex2=1x2+x42x66+x824e^{-x^2} = 1 - x^2 + \dfrac{x^4}{2} - \dfrac{x^6}{6} + \dfrac{x^8}{24} - \cdots

01/2ex2dx=[xx33+x510x742+x9216]01/2\int_0^{1/2}e^{-x^2}\,dx = \left[x - \frac{x^3}{3} + \frac{x^5}{10} - \frac{x^7}{42} + \frac{x^9}{216}\right]_0^{1/2}

=12124+132015376+1110592= \frac{1}{2} - \frac{1}{24} + \frac{1}{320} - \frac{1}{5376} + \frac{1}{110592}

=0.50.041667+0.0031250.000186+0.000009=0.461281= 0.5 - 0.041667 + 0.003125 - 0.000186 + 0.000009 = 0.461281

The error is bounded by the next term: LB1RB◆◆LB11×211RB=1225280.000044\dfrac◆LB◆1◆RB◆◆LB◆11 \times 2^{11}◆RB◆ = \dfrac{1}{22528} \approx 0.000044.

So the integral 0.4613\approx 0.4613 to 4 decimal places.


8. Connections to Other Topics

8.1 Maclaurin series and differential equations

The Maclaurin series provides a method for solving differential equations by substituting a power series ansatz. See Differential Equations.

8.2 Taylor series and complex numbers

Euler's formula eiθ=cosθ+isinθe^{i\theta} = \cos\theta + i\sin\theta is the link between exponential series and trigonometric series. See Complex Numbers.

8.3 Binomial series and further algebra

The binomial expansion is essential for partial fraction decomposition and generating functions. See Further Algebra.


9. Additional Exam-Style Questions

Question 11

(a) Find the Maclaurin series of ln ⁣(1+x1x)\ln\!\left(\dfrac{1 + x}{1 - x}\right) up to the x5x^5 term.

(b) State the range of validity.

Solution

(a) ln ⁣(1+x1x)=ln(1+x)ln(1x)\ln\!\left(\dfrac{1+x}{1-x}\right) = \ln(1+x) - \ln(1-x)

ln(1+x)=xx22+x33x44+x55\ln(1+x) = x - \dfrac{x^2}{2} + \dfrac{x^3}{3} - \dfrac{x^4}{4} + \dfrac{x^5}{5} - \cdots

ln(1x)=xx22x33x44x55\ln(1-x) = -x - \dfrac{x^2}{2} - \dfrac{x^3}{3} - \dfrac{x^4}{4} - \dfrac{x^5}{5} - \cdots

ln ⁣(1+x1x)=2x+2x33+2x55+\ln\!\left(\frac{1+x}{1-x}\right) = 2x + \frac{2x^3}{3} + \frac{2x^5}{5} + \cdots

(b) Valid for 1<x<1-1 < x < 1 (intersection of validity of ln(1+x)\ln(1+x) and ln(1x)\ln(1-x)).

Question 12

Use series to find limx0LB1cosxx22RB◆◆LBx4RB\displaystyle\lim_{x\to 0}\frac◆LB◆1 - \cos x - \frac{x^2}{2}◆RB◆◆LB◆x^4◆RB◆.

Solution

cosx=1x22+x424\cos x = 1 - \dfrac{x^2}{2} + \dfrac{x^4}{24} - \cdots

LB1cosxx2/2RB◆◆LBx4RB=LB1(1x2/2+x4/24)x2/2RB◆◆LBx4RB=LBx4/24+RB◆◆LBx4RB124\frac◆LB◆1 - \cos x - x^2/2◆RB◆◆LB◆x^4◆RB◆ = \frac◆LB◆1 - (1 - x^2/2 + x^4/24 - \cdots) - x^2/2◆RB◆◆LB◆x^4◆RB◆ = \frac◆LB◆-x^4/24 + \cdots◆RB◆◆LB◆x^4◆RB◆ \to \boxed{-\frac{1}{24}}

Question 13

Prove by induction that dndxn(xex)=(x+n)ex\displaystyle\frac{d^n}{dx^n}(xe^x) = (x + n)e^x for all n0n \geq 0.

Solution

Base case (n=0n = 0): d0dx0(xex)=xex=(x+0)ex\dfrac{d^0}{dx^0}(xe^x) = xe^x = (x+0)e^x. True.

Inductive step. Assume dkdxk(xex)=(x+k)ex\dfrac{d^k}{dx^k}(xe^x) = (x+k)e^x.

dk+1dxk+1(xex)=ddx[(x+k)ex]=ex+(x+k)ex=(x+k+1)ex\frac{d^{k+1}}{dx^{k+1}}(xe^x) = \frac{d}{dx}[(x+k)e^x] = e^x + (x+k)e^x = (x+k+1)e^x \quad \blacksquare

Question 14

Find the Maclaurin series of LBcosxRB\sqrt◆LB◆\cos x◆RB◆ up to the x4x^4 term.

Solution

cosx=1x22+x424\cos x = 1 - \dfrac{x^2}{2} + \dfrac{x^4}{24} - \cdots

LBcosxRB=(1x2/2+x4/24)1/2\sqrt◆LB◆\cos x◆RB◆ = (1 - x^2/2 + x^4/24 - \cdots)^{1/2}

Using (1+u)1/2=1+u/2u2/8+(1+u)^{1/2} = 1 + u/2 - u^2/8 + \cdots with u=x2/2+x4/24u = -x^2/2 + x^4/24:

u2=x4/4+u^2 = x^4/4 + \cdots

LBcosxRB=1+12 ⁣(x22+x424)18 ⁣(x44)+\sqrt◆LB◆\cos x◆RB◆ = 1 + \frac{1}{2}\!\left(-\frac{x^2}{2} + \frac{x^4}{24}\right) - \frac{1}{8}\!\left(\frac{x^4}{4}\right) + \cdots

=1x24+x448x432+=1x24x496+= 1 - \frac{x^2}{4} + \frac{x^4}{48} - \frac{x^4}{32} + \cdots = 1 - \frac{x^2}{4} - \frac{x^4}{96} + \cdots

Question 15

Use the Maclaurin series for arctanx\arctan x to show that 113+1517+=LBπRB◆◆LB4RB\displaystyle 1 - \frac{1}{3} + \frac{1}{5} - \frac{1}{7} + \cdots = \frac◆LB◆\pi◆RB◆◆LB◆4◆RB◆.

Solution

arctanx=xx33+x55x77+\arctan x = x - \dfrac{x^3}{3} + \dfrac{x^5}{5} - \dfrac{x^7}{7} + \cdots for x1|x| \leq 1.

Setting x=1x = 1:

arctan1=113+1517+=LBπRB◆◆LB4RB\arctan 1 = 1 - \frac{1}{3} + \frac{1}{5} - \frac{1}{7} + \cdots = \frac◆LB◆\pi◆RB◆◆LB◆4◆RB◆

Therefore n=0(1)n2n+1=LBπRB◆◆LB4RB\displaystyle\sum_{n=0}^{\infty}\frac{(-1)^n}{2n+1} = \frac◆LB◆\pi◆RB◆◆LB◆4◆RB◆. \blacksquare


10. Advanced Worked Examples

Example 10.1: Maclaurin series of exsinxe^x\sin x

Problem. Find the Maclaurin series of exsinxe^x\sin x up to the term in x5x^5.

Solution. ex=1+x+x22+x36+x424+x5120+e^x = 1+x+\dfrac{x^2}{2}+\dfrac{x^3}{6}+\dfrac{x^4}{24}+\dfrac{x^5}{120}+\cdots

sinx=xx36+x5120\sin x = x - \dfrac{x^3}{6} + \dfrac{x^5}{120} - \cdots

Multiplying and collecting terms up to x5x^5:

exsinx=(1+x+x22+x36+x424+x5120) ⁣(xx36+x5120)e^x\sin x = \left(1+x+\frac{x^2}{2}+\frac{x^3}{6}+\frac{x^4}{24}+\frac{x^5}{120}\right)\!\left(x-\frac{x^3}{6}+\frac{x^5}{120}\right)

=xx36+x5120+x2x46+x5120+x32x512+x46x572+x524x6144+= x - \frac{x^3}{6} + \frac{x^5}{120} + x^2 - \frac{x^4}{6} + \frac{x^5}{120} + \frac{x^3}{2} - \frac{x^5}{12} + \frac{x^4}{6} - \frac{x^5}{72} + \frac{x^5}{24} - \frac{x^6}{144} + \cdots

=x+x2+x33x530+O(x6)= x + x^2 + \dfrac{x^3}{3} - \boxed{\dfrac{x^5}{30} + O(x^6)}

Wait, let me be more careful:

x5x^5 coefficient: 1120+1120112172+124\dfrac{1}{120} + \dfrac{1}{120} - \dfrac{1}{12} - \dfrac{1}{72} + \dfrac{1}{24}

=6+6305+15720=8720=190= \dfrac{6 + 6 - 30 - 5 + 15}{720} = \dfrac{-8}{720} = -\dfrac{1}{90}.

So exsinx=x+x2+x33x530+e^x\sin x = x + x^2 + \dfrac{x^3}{3} - \dfrac{x^5}{30} + \cdots

Let me recompute the x4x^4 and x5x^5 terms carefully:

x4x^4: from (16x4)(1)+(x4/24)(x/x)...(-\frac{1}{6}x^4)(1) + (x^4/24)(x/x)... Actually:

  • 1×(x36)1 \times (-\frac{x^3}{6}) contributes nothing to x4x^4
  • x×(x36)=x46x \times (-\frac{x^3}{6}) = -\frac{x^4}{6}
  • x22×x=x32\frac{x^2}{2} \times x = \frac{x^3}{2} (not x4x^4)

x4x^4 terms: x(x36)=x46x \cdot (-\frac{x^3}{6}) = -\frac{x^4}{6} and x424x\frac{x^4}{24} \cdot x... no, that's x5x^5.

Actually x4x^4 comes from: x(x36)=x46x \cdot (-\frac{x^3}{6}) = -\frac{x^4}{6}.

So coefficient of x4x^4 is 16-\dfrac{1}{6}.

exsinx=x+x2+x33x46x530+O(x6)\boxed{e^x\sin x = x + x^2 + \frac{x^3}{3} - \frac{x^4}{6} - \frac{x^5}{30} + O(x^6)}

Example 10.2: Taylor series about a non-zero point

Problem. Find the Taylor series of lnx\ln x about x=1x = 1 up to the term in (x1)4(x-1)^4.

Solution. f(x)=lnxf(x) = \ln x, f(1)=0f(1) = 0.

f(x)=1xf'(x) = \dfrac{1}{x}, f(1)=1f'(1) = 1. f(x)=1x2f''(x) = -\dfrac{1}{x^2}, f(1)=1f''(1) = -1. f(x)=2x3f'''(x) = \dfrac{2}{x^3}, f(1)=2f'''(1) = 2. f(4)(x)=6x4f^{(4)}(x) = -\dfrac{6}{x^4}, f(4)(1)=6f^{(4)}(1) = -6.

lnx=(x1)(x1)22+(x1)33(x1)44+\ln x = (x-1) - \frac{(x-1)^2}{2} + \frac{(x-1)^3}{3} - \frac{(x-1)^4}{4} + \cdots

Example 10.3: Using series to evaluate a limit

Problem. Find limx0LBex1xx22RB◆◆LBx3RB\displaystyle\lim_{x \to 0} \frac◆LB◆e^x - 1 - x - \frac{x^2}{2}◆RB◆◆LB◆x^3◆RB◆.

Solution. ex=1+x+x22+x36+x424+e^x = 1 + x + \dfrac{x^2}{2} + \dfrac{x^3}{6} + \dfrac{x^4}{24} + \cdots

ex1xx2/2x3=LBx3/6+x4/24+RB◆◆LBx3RB=16+x24+\frac{e^x - 1 - x - x^2/2}{x^3} = \frac◆LB◆x^3/6 + x^4/24 + \cdots◆RB◆◆LB◆x^3◆RB◆ = \frac{1}{6} + \frac{x}{24} + \cdots

limx0=16\boxed{\lim_{x \to 0} = \frac{1}{6}}

Example 10.4: Convergence of a Maclaurin series

Problem. Find the radius of convergence of the Maclaurin series of ln(1+x)\ln(1+x).

Solution. ln(1+x)=n=1(1)n+1xnn\ln(1+x) = \displaystyle\sum_{n=1}^{\infty} \frac{(-1)^{n+1}x^n}{n}.

Ratio test: an+1an=nn+1xx\left|\dfrac{a_{n+1}}{a_n}\right| = \dfrac{n}{n+1}|x| \to |x|.

Converges when x<1|x| < 1. At x=1x = 1: alternating harmonic series (converges). At x=1x = -1: 1n-\sum \dfrac{1}{n} (diverges).

Radius of convergence: 1\boxed{1}, interval: (1,1](-1, 1].

Example 10.5: Series solution to estimate a definite integral

Problem. Use a Maclaurin series to estimate 00.5ex2dx\displaystyle\int_0^{0.5} e^{-x^2}\,dx to 4 decimal places.

Solution. ex2=1x2+x42x66+x824e^{-x^2} = 1 - x^2 + \dfrac{x^4}{2} - \dfrac{x^6}{6} + \dfrac{x^8}{24} - \cdots

00.5ex2dx=[xx33+x510x742+x9216]00.5\int_0^{0.5} e^{-x^2}\,dx = \left[x - \frac{x^3}{3} + \frac{x^5}{10} - \frac{x^7}{42} + \frac{x^9}{216} - \cdots\right]_0^{0.5}

=0.50.1253+0.03125100.007812542+0.001953125216= 0.5 - \dfrac{0.125}{3} + \dfrac{0.03125}{10} - \dfrac{0.0078125}{42} + \dfrac{0.001953125}{216}

=0.50.04167+0.0031250.000186+0.000009=0.4613= 0.5 - 0.04167 + 0.003125 - 0.000186 + 0.000009 = \boxed{0.4613} (4 d.p.)

Example 10.6: Differentiating a known series

Problem. By differentiating the Maclaurin series for 11x\dfrac{1}{1-x}, find the Maclaurin series for 1(1x)2\dfrac{1}{(1-x)^2}.

Solution. 11x=n=0xn=1+x+x2+x3+\dfrac{1}{1-x} = \sum_{n=0}^{\infty} x^n = 1 + x + x^2 + x^3 + \cdots

Differentiating: 1(1x)2=n=1nxn1=n=0(n+1)xn=1+2x+3x2+4x3+\dfrac{1}{(1-x)^2} = \sum_{n=1}^{\infty} nx^{n-1} = \sum_{n=0}^{\infty} (n+1)x^n = 1 + 2x + 3x^2 + 4x^3 + \cdots

Example 10.7: Multiplying two Taylor series

Problem. Find the Maclaurin series of cosxex\cos x \cdot e^{-x} up to x4x^4.

Solution. cosx=1x22+x424\cos x = 1 - \dfrac{x^2}{2} + \dfrac{x^4}{24} - \cdots

ex=1x+x22x36+x424e^{-x} = 1 - x + \dfrac{x^2}{2} - \dfrac{x^3}{6} + \dfrac{x^4}{24} - \cdots

Multiplying: (1x22+x424)(1x+x22x36+x424)(1 - \frac{x^2}{2} + \frac{x^4}{24})(1 - x + \frac{x^2}{2} - \frac{x^3}{6} + \frac{x^4}{24})

=1x+x22x36+x424x22+x32x44+x424+= 1 - x + \frac{x^2}{2} - \frac{x^3}{6} + \frac{x^4}{24} - \frac{x^2}{2} + \frac{x^3}{2} - \frac{x^4}{4} + \frac{x^4}{24} + \cdots

=1x+0x33+(12414+124)x4+= 1 - x + 0 - \frac{x^3}{3} + \left(\frac{1}{24}-\frac{1}{4}+\frac{1}{24}\right)x^4 + \cdots

Wait: 124624+124=424=16\frac{1}{24}-\frac{6}{24}+\frac{1}{24} = -\frac{4}{24} = -\frac{1}{6}.

cosxex=1xx33x46+O(x5)\boxed{\cos x \cdot e^{-x} = 1 - x - \frac{x^3}{3} - \frac{x^4}{6} + O(x^5)}


11. Common Pitfalls

PitfallCorrect Approach
Confusing the Maclaurin series (about x=0x=0) with a general Taylor seriesMaclaurin: a=0a=0; Taylor: f(x)=f(n)(a)n!(xa)nf(x) = \sum \dfrac{f^{(n)}(a)}{n!}(x-a)^n
Forgetting the factorial in the denominatorEach term has f(n)(0)n!xn\dfrac{f^{(n)}(0)}{n!}x^n, not f(n)(0)nxn\dfrac{f^{(n)}(0)}{n}x^n
Using a series outside its radius of convergenceAlways check: e.g., ln(1+x)\ln(1+x) converges only for 1<x1-1 < x \leq 1
Incorrectly multiplying seriesCollect like powers carefully; use a table if needed

12. Additional Exam-Style Questions

Question 8

Find the Maclaurin series of (1+x)1/2(1+x)^{-1/2} and determine its radius of convergence.

Solution

(1+x)1/2=1x2+3x285x316+35x4128(1+x)^{-1/2} = 1 - \dfrac{x}{2} + \dfrac{3x^2}{8} - \dfrac{5x^3}{16} + \dfrac{35x^4}{128} - \cdots

Using the general binomial: n=0(1/2n)xn\displaystyle\sum_{n=0}^{\infty} \binom{-1/2}{n} x^n.

Radius of convergence: x<1|x| < 1 (from the binomial series convergence condition).

Question 9

Prove that 01LBln(1+x)RB◆◆LBxRBdx=LBπ2RB◆◆LB12RB\displaystyle\int_0^1 \frac◆LB◆\ln(1+x)◆RB◆◆LB◆x◆RB◆\,dx = \frac◆LB◆\pi^2◆RB◆◆LB◆12◆RB◆.

Solution

ln(1+x)=n=1(1)n+1xnn\ln(1+x) = \displaystyle\sum_{n=1}^{\infty} \frac{(-1)^{n+1}x^n}{n} for x<1|x| < 1.

LBln(1+x)RB◆◆LBxRB=n=1(1)n+1xn1n\dfrac◆LB◆\ln(1+x)◆RB◆◆LB◆x◆RB◆ = \displaystyle\sum_{n=1}^{\infty} \frac{(-1)^{n+1}x^{n-1}}{n}.

01LBln(1+x)RB◆◆LBxRBdx=n=1(1)n+1n2=η(2)=LBπ2RB◆◆LB12RB\displaystyle\int_0^1 \frac◆LB◆\ln(1+x)◆RB◆◆LB◆x◆RB◆\,dx = \sum_{n=1}^{\infty} \frac{(-1)^{n+1}}{n^2} = \eta(2) = \frac◆LB◆\pi^2◆RB◆◆LB◆12◆RB◆.

This is the Dirichlet eta function evaluated at 2. \blacksquare

Question 10

Use the Maclaurin series for sinx\sin x to find sin0.1\sin 0.1 correct to 8 decimal places.

Solution

sin0.1=0.10.0016+0.00001120=0.10.00016667+0.00000008\sin 0.1 = 0.1 - \dfrac{0.001}{6} + \dfrac{0.00001}{120} - \cdots = 0.1 - 0.00016667 + 0.00000008 - \cdots

sin0.10.09983342\boxed{\sin 0.1 \approx 0.09983342}

The x7x^7 term contributes approximately 101210^{-12}, which is negligible.


13. Advanced Topics

13.1 Lagrange form of the remainder

The error in truncating a Taylor series after nn terms is:

Rn(x)=f(n+1)(c)(n+1)!(xa)n+1R_n(x) = \frac{f^{(n+1)}(c)}{(n+1)!}(x-a)^{n+1}

for some cc between aa and xx. This gives a bound on the truncation error.

13.2 Power series solutions of ODEs

The Maclaurin series method can solve ODEs that cannot be solved by standard methods. Substitute y=anxny = \sum a_n x^n into the ODE and equate coefficients.

13.3 Standard Maclaurin series (reference)

FunctionSeriesRadius
exe^xn=0xnn!\displaystyle\sum_{n=0}^{\infty} \frac{x^n}{n!}\infty
sinx\sin xn=0(1)nx2n+1(2n+1)!\displaystyle\sum_{n=0}^{\infty} \frac{(-1)^n x^{2n+1}}{(2n+1)!}\infty
cosx\cos xn=0(1)nx2n(2n)!\displaystyle\sum_{n=0}^{\infty} \frac{(-1)^n x^{2n}}{(2n)!}\infty
ln(1+x)\ln(1+x)n=1(1)n+1xnn\displaystyle\sum_{n=1}^{\infty} \frac{(-1)^{n+1}x^n}{n}11
(1+x)α(1+x)^\alphan=0(L)BαRB◆◆LBnRBxn\displaystyle\sum_{n=0}^{\infty} \binom◆LB◆\alpha◆RB◆◆LB◆n◆RB◆x^n11
arctanx\arctan xn=0(1)nx2n+12n+1\displaystyle\sum_{n=0}^{\infty} \frac{(-1)^n x^{2n+1}}{2n+1}11
11x\dfrac{1}{1-x}n=0xn\displaystyle\sum_{n=0}^{\infty} x^n11

13.4 Using series to prove identities

Many trigonometric identities can be derived from series. For example, eix=cosx+isinxe^{ix} = \cos x + i\sin x implies all the standard addition formulae.


14. Further Exam-Style Questions

Question 11

Find the Maclaurin series of LBsinxRB◆◆LBxRB\dfrac◆LB◆\sin x◆RB◆◆LB◆x◆RB◆ and determine limx0LBsinxRB◆◆LBxRB\displaystyle\lim_{x \to 0} \frac◆LB◆\sin x◆RB◆◆LB◆x◆RB◆.

Solution

sinx=xx36+x5120\sin x = x - \dfrac{x^3}{6} + \dfrac{x^5}{120} - \cdots

LBsinxRB◆◆LBxRB=1x26+x4120\dfrac◆LB◆\sin x◆RB◆◆LB◆x◆RB◆ = 1 - \dfrac{x^2}{6} + \dfrac{x^4}{120} - \cdots

limx0LBsinxRB◆◆LBxRB=1\displaystyle\lim_{x \to 0} \frac◆LB◆\sin x◆RB◆◆LB◆x◆RB◆ = 1 (the constant term).

Question 12

Prove that the Maclaurin series of cosx\cos x converges to cosx\cos x for all real xx.

Solution

The nn-th derivative of cosx\cos x is one of ±cosx\pm\cos x or ±sinx\pm\sin x, so f(n)(c)1|f^{(n)}(c)| \leq 1 for all cc and nn.

By the Lagrange remainder: Rn(x)LBxn+1RB◆◆LB(n+1)!RB|R_n(x)| \leq \dfrac◆LB◆|x|^{n+1}◆RB◆◆LB◆(n+1)!◆RB◆.

For any fixed xx: limnLBxn+1RB◆◆LB(n+1)!RB=0\displaystyle\lim_{n \to \infty} \frac◆LB◆|x|^{n+1}◆RB◆◆LB◆(n+1)!◆RB◆ = 0 (factorial grows faster than exponential).

Therefore Rn(x)0R_n(x) \to 0 and the series converges to cosx\cos x. \blacksquare

Question 13

Use series to evaluate limx0LBtanxxRB◆◆LBx3RB\displaystyle\lim_{x \to 0} \frac◆LB◆\tan x - x◆RB◆◆LB◆x^3◆RB◆.

Solution

tanx=x+x33+2x515+\tan x = x + \dfrac{x^3}{3} + \dfrac{2x^5}{15} + \cdots

LBtanxxRB◆◆LBx3RB=LBx3/3+2x5/15+RB◆◆LBx3RB=13+2x215+\dfrac◆LB◆\tan x - x◆RB◆◆LB◆x^3◆RB◆ = \dfrac◆LB◆x^3/3 + 2x^5/15 + \cdots◆RB◆◆LB◆x^3◆RB◆ = \dfrac{1}{3} + \dfrac{2x^2}{15} + \cdots

limx0LBtanxxRB◆◆LBx3RB=13\boxed{\displaystyle\lim_{x \to 0} \frac◆LB◆\tan x - x◆RB◆◆LB◆x^3◆RB◆ = \frac{1}{3}}


15. Advanced Topics

15.1 Power series solutions of differential equations

For an ODE y+p(x)y+q(x)y=0y'' + p(x)y' + q(x)y = 0, assume y=n=0anxny = \sum_{n=0}^{\infty} a_n x^n.

Substitute into the ODE and equate coefficients of each power of xx to zero. This gives a recurrence relation for ana_n.

15.2 The exponential generating function

E(x)=n=0anxnn!E(x) = \sum_{n=0}^{\infty} \dfrac{a_n x^n}{n!}.

This is useful in combinatorics and probability (e.g., the exponential generating function of the Bernoulli numbers).

15.3 Binomial series — convergence at the endpoints

(1+x)α(1+x)^\alpha converges at x=1x = 1 when α>1\alpha > -1 and at x=1x = -1 when α>0\alpha > 0.

Example: (1+x)1/2(1+x)^{1/2} converges at x=1x = 1 (giving 2\sqrt{2}) but diverges at x=1x = -1.

15.4 Series acceleration

Techniques like Euler acceleration or Shanks transformation can speed up the convergence of slowly converging alternating series.


16. Further Exam-Style Questions

Question 14

Find the Maclaurin series of ln(1x2)\ln(1-x^2) up to x6x^6 and state the radius of convergence.

Solution

ln(1u)=n=1unn\ln(1-u) = -\displaystyle\sum_{n=1}^{\infty} \frac{u^n}{n} for u<1|u| < 1. With u=x2u = x^2:

ln(1x2)=n=1x2nn=x2x42x63\ln(1-x^2) = -\displaystyle\sum_{n=1}^{\infty} \frac{x^{2n}}{n} = -x^2 - \dfrac{x^4}{2} - \dfrac{x^6}{3} - \cdots

Radius of convergence: x2<1    x<1|x^2| < 1 \implies |x| < 1. At x=1x = 1: 1n-\sum \dfrac{1}{n} diverges.

ln(1x2)=x2x42x63+O(x8)\boxed{\ln(1-x^2) = -x^2 - \dfrac{x^4}{2} - \dfrac{x^6}{3} + O(x^8)}, radius =1= 1.

Question 15

Prove that n=1n2n=2\displaystyle\sum_{n=1}^{\infty} \frac{n}{2^n} = 2.

Solution

S=n=1n2n=n=1nxnx=1/2S = \displaystyle\sum_{n=1}^{\infty} \frac{n}{2^n} = \sum_{n=1}^{\infty} n x^n\Big|_{x=1/2}.

We know n=1nxn1=1(1x)2\displaystyle\sum_{n=1}^{\infty} n x^{n-1} = \frac{1}{(1-x)^2} for x<1|x| < 1.

So n=1nxn=x(1x)2\displaystyle\sum_{n=1}^{\infty} n x^n = \frac{x}{(1-x)^2}.

At x=1/2x = 1/2: S=1/2(1/2)2=1/21/4=2S = \dfrac{1/2}{(1/2)^2} = \dfrac{1/2}{1/4} = 2. \blacksquare